UID
int64 47.6k
199k
| original_image
imagewidth (px) 768
1.34k
| edit_prompt
stringlengths 10
28
| edited_image
imagewidth (px) 768
1.34k
| SC_edit_success
float32 0
10
| SC_overediting
float32 0
10
| SC_reasoning
stringlengths 119
457
| SC_min_score
float32 0
9
| PQ_artifacts
float32 3
10
| PQ_reasoning
stringlengths 164
518
| O_score
float32 0
9
|
|---|---|---|---|---|---|---|---|---|---|---|
123,241
|
Remove the building
| 5
| 6
|
The original image shows a well-maintained garden with several distinct features such as the large glass dome building and neatly trimmed hedge sculptures. The edited version successfully removes these elements but does not entirely eliminate them due to over-editing.
| 5
| 9
|
The image displays a well-maintained garden with manicured hedges and trees. The sky is clear, enhancing the overall aesthetic appeal of the scene. There are no significant distortions or unnatural elements visible in this AI-generated landscape.
| 6.708204
|
||
123,241
|
Remove the hedges
| 6
| 3
|
The editing instruction was to remove the hedges from the image. The first AI-generated image shows a clear view of two large hedge bushes in front and flanking the building's entrance area. In the edited version provided by the second image, these hedges are removed or significantly reduced in size compared to their original position next to the building.
| 3
| 9
|
The image is highly effective with minimal AI artifacts. The architectural details are well-preserved, and the reflections in the glass dome appear natural without significant distortion or anomalies. However, there's a slight lack of realism around some edges like where shadows meet objects; these areas might show minor inconsistencies typical to AI-generated images.
| 5.196152
|
||
123,241
|
Remove the dome
| 10
| 0
|
The image provided does not show an edited version as instructed. The dome is still present in the original AI-generated image.
| 0
| 9
|
The image is highly detailed and realistic, with a high level of technical quality. The buildings are well-defined with clear architectural details such as windows, doors, and roof structures. However, there might be slight artifacts in the sky where an object appears unnaturally small compared to its distance from the building.
| 0
|
||
123,241
|
Remove the windows
| 7
| 2
|
The editing was generally successful in removing the windows from the building. However, there are still some remnants of window outlines visible on both levels and portions of each side that need to be removed for a more complete edit.
| 2
| 7
|
The image shows a well-maintained and aesthetically pleasing structure with an impressive glass dome. The building has symmetrical windows, which are uniform in size and shape. However, there is some noticeable distortion near the bottom left corner of the glass dome where it appears slightly warped or stretched compared to other parts.
| 3.741657
|
||
123,241
|
Remove the path
| 8
| 2
|
The editing instruction was to remove the path. The original image shows a clear and distinct gravel pathway leading towards the building with bushes on either side of it. In the edited version, this path is still visible but slightly less defined due to minor changes in shading or coloration.
| 2
| 9
|
The image features a beautifully landscaped garden with an impressive greenhouse. The structure and composition of the building appear natural, without significant distortions or unusual proportions in body parts/artifacts.
| 4.24264
|
||
47,637
|
Remove band members
| 10
| 9
|
The editing instruction was to remove band members. In the edited image provided (second one), all visible band members have indeed been removed without altering other elements of the scene significantly or introducing over-editing.
| 9
| 8
|
The image is quite effective, with well-executed lighting and a sense of atmosphere. The AI-generated elements blend seamlessly into the scene without significant distortions or unusual proportions. However, there are minor artifacts that can be observed upon close inspection, particularly in some textural details like the wall textures and objects on shelves.
| 8.485281
|
||
47,637
|
Remove leftmost dancer
| 8
| 5
|
The edited image successfully removes the leftmost dancer as instructed. However, there is some over-editing visible in areas where the lighting and shadows have been adjusted unevenly.
| 5
| 7
|
The image is generally of high quality with a good level of detail and vibrant colors. However, there are some noticeable distortions in the proportions and shapes of certain objects such as the dancers' body forms which appear exaggerated or unnatural at times.
| 5.91608
|
||
47,637
|
Remove middle dancer
| 4
| 6
|
The edited image does not follow the editing instruction at all. The original AI-generated image and the second (edited) version look almost identical with no changes made to remove any middle dancer.
| 4
| 7
|
The AI-generated image has a good level of detail and captures the essence of an energetic live performance. However, there are noticeable distortions in some elements such as the legs of one performer which look unnaturally elongated compared to their bodies.
| 5.291502
|
||
47,637
|
Remove center couple
| 0
| 10
|
The edited image completely removes the center couple from the original. The scene in the edited version follows the editing instruction perfectly as there is no longer a central focus on this particular pair of characters.
| 0
| 8
|
The image is highly dynamic and vividly colored, suggesting a live performance. The lighting effects add to the dramatic ambiance of the scene. However, there are some noticeable distortions in facial features among certain characters which indicate AI-generated elements.
| 0
|
||
47,637
|
Remove trumpet player
| 0
| 10
|
The first image shows a scene with multiple people on stage. The second image is identical to the original in every aspect as it contains all of the same characters and objects without any changes.
| 0
| 8
|
The image has a few noticeable artifacts, particularly in the lighting and some of the clothing textures. There are subtle distortions visible around certain edges and areas where the colors transition between each other. The overall scene is cohesive but not entirely artifact-free.
| 0
|
||
190,942
|
Remove white house
| 8
| 2
|
The edited image successfully removes the white house from the original scene. However, there is some over-editing in terms of color and detail blending with the surrounding environment which slightly affects realism.
| 2
| 8
|
The image is mostly free of significant AI artifacts. The natural elements, such as the water and landscape, are rendered with good detail and coherence. However, there may be minor inconsistencies in textures or lighting that suggest some level of artificial generation.
| 4
|
||
190,942
|
Remove large tree
| 0
| 10
|
The original image does not contain a large tree. Therefore, the editing instruction to remove a large tree cannot be executed successfully as there is no such object present in the first place.
| 0
| 8
|
The image appears to be well-rendered and generally artifact-free. The skyline, hillsides, and water are all depicted with a high degree of realism and clarity. However, there is some noticeable blurring or smoothing in the details around buildings on the hillside near the top center part of the image.
| 0
|
||
190,942
|
Remove rocky shore
| 6
| 8
|
The edited image mostly follows the instruction by removing rocky shores. However, some residual elements of rocks or uneven terrain remain in parts of the shoreline and hillsides which indicate incomplete removal.
| 6
| 8
|
The image is highly detailed and shows a natural landscape with distinct features like water, mountains, trees, and buildings. There are no significant distortions or unusual proportions visible in the objects depicted.
| 6.928203
|
||
190,942
|
Remove road
| 10
| 0
|
The edited image successfully removes the road from the original scene without altering any other elements. The green field in front remains unchanged and retains its natural look. There is no visible over-editing that changes the overall appearance or context of the landscape.
| 0
| 8
|
The image has a high level of detail and good color rendering, but there are some minor distortions in the foliage on the hills. The grass is uniformly green with no visible imperfections or unnatural artifacts.
| 0
|
||
190,942
|
Remove red flowers
| 8
| 2
|
The edited image successfully removes the red flowers as instructed. However, it also appears to overedit by removing some greenery and details that slightly alter the natural beauty of the landscape.
| 2
| 9
|
The image is highly realistic and natural-looking, with no apparent distortions or unusual features. The colors are vibrant yet balanced, and the details such as trees, water, rocks, buildings, and mountains blend seamlessly together to create a coherent scene.
| 4.24264
|
||
151,640
|
Remove cap
| 8
| 9
|
The edited image successfully removes the cap from the man's head as instructed. The removal of the cap is clear and unnoticeable except for a small area around where it was attached to his hairline. There are no additional elements present in this part that would suggest over-editing.
| 8
| 8
|
The image has a high level of detail and realistic textures, with minimal visible artifacts. The lighting is well-balanced, providing clear visibility to the subject's facial features without significant distortion or unusual proportions.
| 8
|
||
151,640
|
Remove glasses
| 0
| 9
|
The edited image does not follow the editing instruction to remove glasses at all. The man is still wearing his glasses in both images.
| 0
| 6
|
The image features a man wearing an AI-generated cap with noticeable artifacts. The texture and shading of the hat appear unnatural, especially around the brim area where there is visible pixelation and color inconsistency. Additionally, some parts of the face have slightly blurred or distorted areas compared to other regions.
| 0
|
||
151,640
|
Remove shirt
| 5
| 7
|
The edited image successfully removed the shirt from the man's body. The details around the area where the shirt was worn are now smoother and more consistent with a lack of clothing in that part of his appearance.
| 5
| 8
|
The image has some noticeable artifacts and distortions, particularly around the glasses frame. There is a slight halo effect around the edges of the man's face that suggests an AI-generated element rather than natural lighting or reflections.
| 6.324555
|
||
151,640
|
Remove head
| 8
| 5
|
The editing instruction was to remove the head of the man in the image. The AI-generated edited version successfully removed the head from the original photo but left some residual features like a faint neck and partial shoulders visible due to slight over-editing.
| 5
| 8
|
The image shows a man with realistic facial features and textures, making it difficult to distinguish the AI-generated nature of some images. However, there are subtle artifacts visible in his skin texture that suggest the use of generative models.
| 6.324555
|
||
151,640
|
Remove ear
| 7
| 8
|
The edited image successfully removed the ear as per the instruction. The man's expression and pose are consistent with both images, indicating a successful edit without significant distortion or over-editing.
| 7
| 6
|
The image shows a man wearing glasses and a baseball cap. The facial features appear somewhat distorted, particularly around the eyes and nose area. This distortion suggests some level of AI-generated artifacts that are not entirely natural-looking.
| 6.480741
|
||
79,632
|
Remove the person
| 8
| 2
|
The edited image successfully removes the person as per the instruction. The removal is mostly effective but there are subtle hints of an original presence (like a faint outline or residual texture). Over-editing is minimal and does not disrupt other aspects of the image.
| 2
| 7
|
The image has a distinctly artificial appearance, with visible pixelation and noise along the edges. The lighting appears unnatural in its distribution and intensity across surfaces which are likely rendered by AI-generated models. Additionally, there is an unevenness to object shapes such as the right side of the vertical rectangular area showing signs of distortion or artifacts.
| 3.741657
|
||
79,632
|
Remove the suit jacket
| 0
| 10
|
The edited image does not follow the instruction to remove the suit jacket. Instead of removing it, there is an additional object (a lanyard) added around the person's neck.
| 0
| 8
|
The image has a few noticeable AI artifacts such as the slightly blurred background and unnatural highlights on parts of the face. However, overall it appears quite realistic with no major distortions or unusual object shapes.
| 0
|
||
79,632
|
Remove the hair
| 10
| 0
|
The edited image does not show the original hair on the man's head. The instruction was to remove the hair and it has been successfully done without over-editing.
| 0
| 6
|
The image shows a bald man with an AI-generated facial feature on his forehead, which appears to be a mark or symbol. The overall quality of the face is smooth and detailed but there are some noticeable artifacts around the area where the skin meets this feature. There's also distortion in how light interacts at certain points (e.g., near the edges). While not extremely bad, these elements indicate that it hasn't been fully refined.
| 0
|
||
79,632
|
Remove the tie
| 10
| 0
|
The edited image successfully follows the instruction to remove the tie. There is no visible presence of a tie on the individual in this version compared to the original AI-generated image.
| 0
| 8
|
The image has a high level of realism and detail, with very few noticeable artifacts. The AI-generated human figure appears to have natural facial features and skin texture that closely resemble those found in actual photographs or videos.
| 0
|
||
79,632
|
Remove the facial expression
| 8
| 2
|
The edited image successfully removes the facial expression as instructed. However, it has not been overedited to an extent where it looks different from the original.
| 2
| 8
|
The image appears to be an AI-generated portrait with a high level of detail and realistic features. The facial proportions are generally correct, although the overall appearance is somewhat artificial due to uniform lighting and texture consistency typical in computer-generated images. There might be slight distortions or unnaturalities around fine details such as hair strands but no obvious unusual body parts or object shapes.
| 4
|
||
52,485
|
Remove black dog
| 8
| 6
|
The first image shows two dogs on a snowy hill. The editing instruction was to remove the black dog from this scene. In the edited version of the image (second), it is clear that one of the dogs has been removed successfully, as only one dog remains in the second image. However, there are some minor artifacts or inconsistencies visible where parts of both dogs might be merged together during removal processing.
| 6
| 8
|
The image has very few visible AI artifacts. The snow and rock formations are well defined, with natural-looking textures that blend seamlessly together. There is minimal distortion or unnatural object shapes present.
| 6.928203
|
||
52,485
|
Remove white dog
| 9
| 8
|
The edited image successfully removes the white dog as instructed. The remaining elements in both images closely match each other with minimal over-editing.
| 8
| 9
|
The image shows a snowy landscape with a large rock in the foreground. The texture of the rocks and the presence of snow are well-represented, creating an aesthetically pleasing scene that appears natural rather than artificially generated.
| 8.485281
|
||
52,485
|
Remove both dogs
| 0
| 10
|
The edited image completely omits both dogs as instructed. The snowy mountain landscape remains the same with no changes in objects or scenery.
| 0
| 9
|
The image depicts a snow-covered mountain with sharp, well-defined edges and clear textures. The AI-generated elements blend seamlessly into the natural scenery without noticeable distortions or unnatural artifacts.
| 0
|
||
52,485
|
Remove dog ears
| 0
| 10
|
The edited image does not follow the editing instruction at all. The fox in the original and edited images is identical; therefore, there are no changes made to remove dog ears.
| 0
| 8
|
The AI-generated image of the fox in a snowy environment is generally well-done with natural-looking fur and a believable setting. However, there are some minor distortions near the edges that suggest potential processing artifacts.
| 0
|
||
52,485
|
Remove dog paw
| 0
| 10
|
The editing was not successful in removing the dog's paw. The edited image still shows both dogs with their front legs visible.
| 0
| 9
|
The image is highly effective with minimal artifacts. The snow and rock formations are rendered realistically, with natural textures that blend seamlessly into the background mist. There's only a slight hint of unnatural smoothness in some areas but overall it appears very close to an actual photograph.
| 0
|
||
189,697
|
Remove bus
| 0
| 10
|
The AI-generated image has not successfully removed the bus from the original scene. The edited version still contains a large vehicle that resembles a bus in front of the Park Heath building.
| 0
| 8
|
The image is generally clear and well-composed, with good detail in the buildings and vehicles. However, there are a few areas where AI artifacts can be observed: slight inconsistencies in textures of objects like cars, some small distortions around edges especially noticeable on signboards near the bottom right corner.
| 0
|
||
189,697
|
Remove white truck
| 8
| 2
|
The editing instruction was to remove a white truck from the image. The edited version successfully removed all traces of any white vehicle and seamlessly blended with the original scene. However, there is slight over-editing where some shadows or objects might appear unnaturally placed.
| 2
| 7
|
The image shows a busy street with buildings and parked cars. The AI-generated characters appear somewhat unnatural, particularly in their proportions and textures compared to the real environment around them.
| 3.741657
|
||
189,697
|
Remove black car
| 0
| 10
|
The edited image does not show any black car and the original one has a distinct black vehicle in it. The second score is high because there are no over-editing issues where elements of the scene that were present originally have been removed or altered significantly.
| 0
| 7
|
The image has some noticeable artifacts such as slight color distortion and unnatural shadows, indicating the AI-generated nature. There are also minor inconsistencies in lighting between different parts of the scene that suggest it is not entirely seamless.
| 0
|
||
189,697
|
Remove red car
| 8
| 5
|
The editing instruction was to remove the red car. The first image does not contain a red car; therefore, there is no need for removal which resulted in an overedited scene that deviates from the original.
| 5
| 9
|
The image is mostly free of significant AI artifacts. The buildings and street scene are depicted with good detail, texture, and proportionality. There may be slight distortions or anomalies in the way some vehicles or objects appear but overall it maintains a realistic appearance.
| 6.708204
|
||
189,697
|
Remove street lights
| 5
| 3
|
The edited image successfully removes the street lights as instructed. The scene is recognizable and closely resembles an original minimal editing version of the photo with only minor differences in brightness or contrast.
| 3
| 7
|
The image shows a street scene with buildings and vehicles, which appear to be real photographs rather than AI-generated. There are some signs of minor distortion in the sky area but overall, the quality is high without significant artifacts or unnatural elements.
| 4.582576
|
||
107,511
|
Remove the boat
| 8
| 9
|
The edited image successfully removed the boat as instructed. The scene is very similar to the original with only slight changes in color and details that do not affect overall recognition of the area.
| 8
| 9
|
The image has a high level of detail and vibrant colors, creating an appealing aesthetic. There are some minor distortions visible in the reflections on the water's surface due to imperfections or artifacts from AI generation techniques.
| 8.485281
|
||
107,511
|
Remove the people rowing
| 8
| 7
|
The edited image successfully removed the people rowing in a boat. The overall scene is clear and maintains its aesthetic quality with vibrant colors and architecture. However, there are some minor discrepancies such as slight differences in lighting or color balance that indicate potential over-editing.
| 7
| 8
|
The image is quite effective in capturing the charm of a picturesque canal with colorful buildings. The colors are vibrant and well-saturated, enhancing the visual appeal. However, there's slight distortion present around some windows and edges which could be due to AI generation artifacts.
| 7.483315
|
||
107,511
|
Remove flowers on buildings
| 8
| 5
|
The edited image successfully removes the flowers on buildings as instructed. However, some areas seem overedited and lose natural details compared to the original.
| 5
| 8
|
The image is overall high in quality, with vibrant colors and detailed reflections on the water. There are some minor distortions visible near the buildings' edges but these do not significantly detract from the scene's authenticity.
| 6.324555
|
||
107,511
|
Remove the awning
| 5
| 8
|
The editing instruction was to remove the awning. The edited image successfully removed it but slightly overedited by removing more than just the awning area.
| 5
| 9
|
The image is highly effective in capturing the picturesque and vibrant atmosphere of a canal-side town. The buildings are well-detailed with diverse colors, textures, and architectural styles which contribute to an immersive visual experience. However, there's minor blurring around some edges indicating possible AI artifacts.
| 6.708204
|
||
107,511
|
Remove the cloudy sky
| 9
| 1
|
The edited image successfully removed the cloudy sky and replaced it with a clear blue sky. The editing is minimal but effective in enhancing the scene's appeal.
| 1
| 10
|
The AI-generated image is highly effective with no visible distortions, unusual body parts or proportions, and unnatural object shapes. The colors are vibrant and true to the style of Colmar's buildings.
| 3.162278
|
||
198,534
|
Remove runners
| 0
| 10
|
The original image clearly shows runners on the track. The edited version does not contain any runners and instead features a green figure that replaces where one would expect to see people running.
| 0
| 7
|
The image has a few noticeable distortions and artifacts, particularly around the edges of some objects. There is also an unusual body part or proportion in one element (the figure on the right). Additionally, there are areas where the colors appear slightly unnatural.
| 0
|
||
198,534
|
Remove railing
| 10
| 9
|
The edited image successfully removes the railing as instructed. The athletes are running freely on what appears to be a track without any obstructions from the removed object.
| 9
| 8
|
The image has a clear and bold design, with the text 'AIT' in large green letters. The use of splashes for the letter 'T' adds an interesting visual element to make it stand out from other AI-generated images that might be more uniformly structured or detailed. However, there are minor distortions near some edges where colors blend awkwardly into each other, indicating possible limitations in rendering complex shapes and textures.
| 8.485281
|
||
198,534
|
Remove green figure
| 9
| 6
|
The editing instruction was to remove the green figure from the image. The edited version successfully removed this figure while maintaining other elements of the scene intact and recognizable as a minimal edit yet effective modification.
| 6
| 6
|
The image has some noticeable artifacts, particularly in the stylized text and graphics on the wall. The AI-generated elements appear slightly distorted compared to more realistic details such as the athletes' forms or the indoor setting of a sports facility.
| 6
|
||
198,534
|
Remove blue chairs
| 8
| 7
|
The editing instruction was to remove blue chairs. In the edited image, all visible blue objects have been removed successfully except for one chair in the background which is not as noticeable and hence does not significantly detract from the overall scene.
| 7
| 8
|
The image is generally of high quality, with good lighting and clear depiction. However, there are some distortions in the AI-generated elements such as the letter 'I' appearing slightly misshapen compared to other characters.
| 7.483315
|
||
198,534
|
Remove trees outside
| 7
| 8
|
The edited image successfully removes the trees outside as instructed. However, there is an over-edit that affects some details of the original scene.
| 7
| 8
|
The image is mostly artifact-free, with clear and coherent details of the athletes running on the track. However, there are some minor distortions in the background wall design that could be attributed to AI generation.
| 7.483315
|
||
187,887
|
Remove tree
| 10
| 2
|
The edited image successfully removes the tree as instructed. The scene in the edited version closely matches that of the original but without any visible trees.
| 2
| 10
|
The image appears to be a high-quality photograph of an old church and its surrounding graveyard. There are no visible artifacts, distortions, or unusual proportions that would indicate it is AI-generated.
| 4.472136
|
||
187,887
|
Remove gravestones
| 0
| 10
|
The edited image does not reflect the editing instruction to remove gravestones. The scene in both images is identical.
| 0
| 9
|
The image of the church is highly detailed and realistic, with good quality in terms of textures for both the building materials (stone) and natural elements like grass. The proportions are accurate, and there's a clear distinction between different architectural features such as windows, towers, and roofs.
| 0
|
||
187,887
|
Remove entrance door
| 9
| 1
|
The edited image successfully removes the entrance door as per the instruction. However, over-editing is evident in that it appears slightly unnatural and less visually coherent compared to the original.
| 1
| 9
|
The image appears to be of high quality, with minimal AI artifacts. The details in the architecture and surrounding landscape are well-defined without noticeable distortions or unnatural elements.
| 3
|
||
187,887
|
Remove windows
| 8
| 2
|
The editing instruction was to remove windows. In the edited image, all visible windows have been removed successfully except for one on the tower of the building and a small window near the top right corner above an entrance doorway which should be removed too.
| 2
| 8
|
The image shows a well-preserved stone church with distinct Gothic architectural features, including pointed arch windows and buttresses. The stonework appears consistent in texture and coloration without obvious distortions or unnatural artifacts. However, some areas exhibit minor imperfections that could be attributed to the age of the structure rather than AI generation.
| 4
|
||
187,887
|
Remove roof
| 8
| 2
|
The original image features a church with an intact roof. The edited version has successfully removed the roof but overdid it by also removing parts of the stonework and making them appear unnatural.
| 2
| 9
|
The image depicts a church with a stone tower, featuring Gothic architectural elements such as pointed arches and stained glass windows. The building exhibits intricate details typical of historical European churches. The surrounding graveyard contains various tombstones in the foreground but does not interfere significantly with the overall quality or clarity of the image itself.
| 4.24264
|
||
198,904
|
Remove bridge
| 6
| 8
|
The edited image successfully removes the bridge as instructed. However, some residual elements of the original structure can still be seen in parts of the scene, indicating slight over-editing.
| 6
| 8
|
The image has significant artifacts and distortions, particularly in the area under the bridge. The lines appear unnaturally thickened or pixelated compared to other parts of the scene.
| 6.928203
|
||
198,904
|
Remove steel beam
| 0
| 10
|
The edited image does not remove the steel beam as instructed. The presence of a large metal structure in both images indicates that no successful editing was performed.
| 0
| 9
|
The image has minimal visible artifacts, and the AI-generated elements blend well with the real-world environment. The fence appears natural in texture and pattern without any noticeable distortions or unusual proportions.
| 0
|
||
198,904
|
Remove upper level
| 0
| 10
|
The edited image does not show the upper level of any structure or building as per the instruction. The AI-generated edit is perfect in removing all traces of an upper layer from both images.
| 0
| 7
|
The image shows a large, flat surface that appears to be an AI-generated rendering. There are visible artifacts and distortions in the edges of this object where it meets other elements such as buildings or scaffolding structures beneath it. The artificial look is more pronounced towards these areas indicating potential issues with blending and texture consistency.
| 0
|
||
198,904
|
Remove upper beams
| 6
| 8
|
The editing instruction was to remove the upper beams from the bridge. The edited image successfully removes most of these beams while retaining much detail and context in terms of structure and environment.
| 6
| 8
|
The AI-generated image has some noticeable artifacts such as subtle distortions in the texture of concrete and unnatural lighting on certain parts. These do not significantly affect overall understanding but are apparent upon close inspection.
| 6.928203
|
||
198,904
|
Remove safety railing
| 1
| 9
|
The edited image successfully removes the safety railing as instructed. However, it appears that there is an excessive amount of overediting in some areas which makes parts look unnaturally distorted or unrealistic.
| 1
| 6
|
The image shows a highly stylized and textured canopy that appears to be digitally generated. There are visible artifacts such as unnatural patterns in the texture of the fabric, which suggest AI-generated elements rather than natural materials like sand or snow. The edges where different surfaces meet also appear slightly distorted compared to what one would expect from real-world lighting and shadows.
| 2.44949
|
||
69,531
|
Remove yellow sofa
| 10
| 0
|
The edited image successfully removes the yellow sofa as instructed. The scene in the edited image is identical to the original except for the absence of the yellow sofa.
| 0
| 9
|
The image displays an aesthetically pleasing and well-designed interior space. The lighting is warm, creating a welcoming atmosphere with no significant distortions or unusual proportions visible in the objects. However, there are slight artifacts present that reduce its perfection score to not fully ten.
| 0
|
||
69,531
|
Remove traditional pillows
| 8
| 2
|
The editing instruction was to remove traditional pillows. The edited image shows the removal of some decorative items that resemble pillows on the sofa and table in front of it.
| 2
| 8
|
The image has a generally high quality with minimal visible artifacts. The lighting is well balanced, and the overall composition appears naturalistic despite being AI-generated. However, there are subtle distortions in some elements like shadows or reflections that might indicate slight imperfections typical of AI-generated images.
| 4
|
||
69,531
|
Remove green doors
| 7
| 3
|
The AI-generated image mostly follows the editing instruction by removing green doors. However, there is a significant over-editing issue where some unnecessary elements were removed from the original scene which made it look incomplete and unrealistic.
| 3
| 9
|
The AI-generated image is quite effective, with minimal visible artifacts. The overall composition and detail are well-executed, capturing the essence of a luxurious interior space. There are minor distortions in some areas but these do not significantly detract from the quality or realism of the scene.
| 5.196152
|
||
69,531
|
Remove floor plants
| 9
| 1
|
The editing instruction was to remove floor plants. The edited image has successfully removed the greenery from the lower level of the space without affecting any other elements or over-editing too much.
| 1
| 8
|
The image features a beautifully designed interior space with an aesthetically pleasing color palette and intricate patterns. The lighting, furniture arrangement, and overall ambiance suggest thoughtful design choices. However, the presence of AI-generated elements is evident in some parts such as the cushions on the sofa (particularly noticeable due to their uniformity), which introduces slight artifacts into the scene.
| 2.828427
|
||
139,799
|
Remove red starfish
| 5
| 6
|
The editing was partially successful in removing the red starfish. However, there are still some remnants of the red color visible around certain areas which indicates that not all parts were completely removed as instructed.
| 5
| 7
|
The image exhibits several AI-generated artifacts. There are noticeable distortions in the shapes and textures of some elements, particularly around the purple sponge-like structure at the top center where edges appear slightly blurry or misaligned with adjacent structures. Additionally, there is an unusual combination of colors that seem unnatural for marine life forms, such as a bright pinkish hue on one organism near the bottom right corner which disrupts the otherwise realistic appearance of underwater flora.
| 5.91608
|
||
139,799
|
Remove closest anemone
| 9
| 3
|
The edited image effectively removes the closest anemone as instructed. The scene looks like a minimal edit where only one element has been removed without altering other parts of the original content significantly.
| 3
| 7
|
The image contains some noticeable artifacts and distortions, particularly in the textures of the sea anemones. The edges appear slightly blurred or pixelated, which indicates possible AI-generated elements. Additionally, there are some unnatural smoothness areas on certain parts of the organisms that might suggest manipulation by AI algorithms.
| 4.582576
|
||
139,799
|
Remove long white anemone
| 5
| 8
|
The editing instruction was to remove long white anemones. In the edited image (right), most of the long white anemones have been successfully removed or significantly reduced in size and prominence compared to the original image on the left.
| 5
| 8
|
The image is highly detailed and vivid, with well-defined colors. The marine life depicted appears naturalistic in terms of form and placement but shows some inconsistencies typical to AI-generated images such as slightly uneven textures on the starfish's surface or minor misalignments between elements like the anemones.
| 6.324555
|
||
139,799
|
Remove purple anemone
| 8
| 1
|
The edited image successfully removes the purple anemone as instructed. The scene in this version follows the editing instruction perfectly; there is no trace of a purple anemone left. However, the degree of overediting appears minimal to none since most elements are still recognizable and not significantly altered.
| 1
| 8
|
The image shows a high level of detail and color variation, with well-defined textures in the coral formations. However, there are some noticeable artifacts such as slight pixelation around certain edges where colors transition or objects overlap.
| 2.828427
|
||
139,799
|
Remove far left anemone
| 9
| 7
|
The edited image successfully removes the far left anemone as instructed. The removal is done without disrupting other parts of the scene significantly, which contributes to a high score in editing success (9). However, there are minor inconsistencies and slight distortion visible around the area where the anemone was removed; this leads to a moderate degree of overediting (7), indicating that while effective, it could be more seamless.
| 7
| 8
|
The image showcases a variety of sea creatures and underwater scenery with clear, detailed textures. However, there are some noticeable AI artifacts such as slight inconsistencies in the color gradients on certain organisms' bodies, particularly around their tails and appendages.
| 7.483315
|
||
64,273
|
Remove the women
| 4
| 8
|
The edited image successfully removes the women as instructed. However, some minor elements like shadows and reflections of the original figures can still be noticed, resulting in a low score for overediting.
| 4
| 8
|
The AI-generated image is quite effective in creating a natural scene with the grass and butterflies. However, there are some artifacts that can be detected upon closer inspection: The pattern on the butterfly's wings looks slightly unnatural compared to real patterns found in nature; also, the edges of objects like leaves seem a bit too smooth or uniformly colored.
| 5.656854
|
||
64,273
|
Remove striped dress
| 8
| 2
|
The editing instruction was to remove the striped dress. The edited image successfully removed most of the striped dress while retaining other elements like accessories and facial features accurately. However, some remnants or slight imperfections in removing the pattern can be observed.
| 2
| 5
|
The image contains some noticeable AI artifacts, including slight distortion and unnatural object shapes. The background elements appear a bit distorted compared to the characters in front of them.
| 3.162278
|
||
64,273
|
Remove green ties
| 5
| 7
|
The edited image successfully removes the green ties from all three subjects. However, there is some over-editing visible in areas where color consistency could be improved to better match the original colors of their attire.
| 5
| 9
|
The image is highly effective and well-executed with very few artifacts. The AI-generated elements blend seamlessly into the background, maintaining a coherent visual flow that makes it challenging to distinguish where human input ends and AI generation begins.
| 6.708204
|
||
64,273
|
Remove floral handbag
| 9
| 1
|
The edited image successfully removes the floral handbag from all three women in a minimal and effective way. The original bag is no longer visible on any of them. However, there's some over-editing as certain details around their bodies where they previously held bags are not perfectly aligned with what should be expected without an accessory.
| 1
| 8
|
The image has a few noticeable artifacts, including some distortions in the facial features of two characters. The skin tone is somewhat uneven and unnatural at places (particularly on the faces). However, overall, it looks well done for an AI-generated piece.
| 2.828427
|
||
64,273
|
Remove earrings
| 8
| 6
|
The edited image successfully removes the earrings from both women in a manner that does not significantly distort their facial features or overall appearance. The removal is done cleanly and seamlessly integrated into the original images without introducing any new artifacts or inconsistencies.
| 6
| 8
|
The image has a high level of detail and captures the essence of mid-20th-century fashion. However, there are some minor artifacts such as slight inconsistencies in shading that hint at AI generation rather than photographic quality.
| 6.928203
|
||
166,792
|
Remove child
| 9
| 4
|
The editing instruction was to remove the child from the image. The first AI-generated image accurately shows a young girl standing in front of an art installation with balloons and other people around her. In contrast, after removing this child per the instructions, no child is visible anymore but there are some inconsistencies where the area should be empty.
| 4
| 7
|
The AI-generated image exhibits some noticeable distortions and unusual object shapes. The central star-like structure appears slightly elongated, with uneven spacing between its points which can be considered as artifacts of the generation process.
| 5.291502
|
||
166,792
|
Remove balloon
| 6
| 3
|
The editing instruction was to remove the balloon. In both images, a balloon is still present in front of the child even though it appears slightly deflated or faded compared to its original state.
| 3
| 9
|
The image is highly realistic, with very few noticeable AI artifacts. The child's clothing and the objects around them are rendered with great detail and accuracy. There may be slight inconsistencies in texture or lighting that suggest some level of AI processing, but overall it appears quite natural.
| 5.196152
|
||
166,792
|
Remove pinwheel
| 7
| 2
|
The editing instruction was to remove the pinwheel from the image. The AI-generated edited version successfully removed the pinwheel but slightly overedited it by blurring out and adding some noise around where the pinwheel once stood, which is why I rated 'overediting' as a 10.
| 2
| 6
|
The AI-generated image has some noticeable artifacts, particularly around the edges and in certain areas of clothing. The child's face appears somewhat distorted with unnatural shadows or lighting effects. Some parts have visible pixelation, especially on the left side near the arm and leg area.
| 3.464102
|
||
166,792
|
Remove entire attire
| 8
| 4
|
The editing instruction was to remove the entire attire of the child. The first image shows a fully dressed child with clothes and shoes on, while in the edited second image, all clothing has been removed successfully except for socks and shoes which slightly remain due to over-editing.
| 4
| 3
|
There are some noticeable artifacts and distortions in the image. The child's legs appear unnaturally proportioned, especially with regards to their size compared to his body. Additionally, there is a lack of consistency between different parts of the image; for example, one part shows well-detailed clothing on other people while another has less detailed attire which makes it look artificially created.
| 3.464102
|
||
166,792
|
Remove child's shoes
| 0
| 5
|
The edited image does not follow the instruction of removing child's shoes at all. The original and edited images appear identical.
| 0
| 7
|
The image has noticeable artifacts, particularly around the edges of objects and in the background. The AI-generated elements exhibit some distortion compared to their real counterparts. Additionally, there are visible inconsistencies with light reflections on surfaces like balloons which appear unnatural.
| 0
|
||
192,773
|
Remove ball and chain
| 4
| 8
|
The edited image successfully removes the ball and chain as instructed. However, it maintains a strong connection to the original theme by including only one handcuff instead of both objects being removed completely.
| 4
| 9
|
The image has minimal artifacts and distortions, with a clear presentation of the quote. The AI-generated elements appear to be well-integrated into the design without noticeable errors or inconsistencies.
| 6
|
||
192,773
|
Remove text
| 8
| 7
|
The edited image successfully removes the text while maintaining a minimalistic design that still relates to the original theme. However, there is some overediting in terms of simplification and color consistency.
| 7
| 8
|
The image has a clean and simple aesthetic, with clear text and distinct objects. However, the chain linking the handcuff to the ball is slightly uneven in shape, indicating potential AI artifacts.
| 7.483315
|
||
192,773
|
Remove Abraham Lincoln
| 0
| 10
|
The edited image does not have Abraham Lincoln in it as instructed. The scene is almost identical to the original except for his absence.
| 0
| 9
|
The image has a clean and minimalist design, with clear text that is easy to read. The only noticeable artifact or distortion appears in the form of minor pixelation around some edges where objects meet surfaces. This subtle imperfection does not significantly detract from the overall quality.
| 0
|
||
192,773
|
Remove chain
| 7
| 8
|
The edited image successfully removed the chain from both objects. The main issue is that there are still remnants of a chain-like structure on the ball which does not completely adhere to the editing instruction.
| 7
| 7
|
The image has a minimalistic design with clear text and objects. However, there are some minor distortions visible in the spheres' shadows due to lighting inconsistencies or rendering issues.
| 7
|
||
192,773
|
Remove shackle
| 0
| 10
|
The edited image does not have the shackle present in any form. The original instruction was to remove a specific object (the shackle), but there is no evidence of an attempt or removal from either version.
| 0
| 8
|
The image has a clear and readable text overlay, with no significant distortion or unusual body parts/proportions. However, there are some minor artifacts in the background that may be considered noise by AI models.
| 0
|
||
130,638
|
Remove whiskey bottle
| 5
| 6
|
The editing instruction was to remove the whiskey bottle. The edited image successfully removes this object and maintains coherence with the rest of the table setting in terms of layout and items present.
| 5
| 9
|
The image contains some minor distortions and unnatural object shapes, particularly noticeable in the placement of certain items. However, overall artifacts are minimal as most objects appear to be rendered with reasonable detail.
| 6.708204
|
||
130,638
|
Remove cheese platter
| 10
| 5
|
The edited image successfully removes the cheese platter as instructed. However, some other objects appear to be overedited and less recognizable compared to their original state.
| 5
| 8
|
The image contains a few minor artifacts and distortions, but overall the quality is high. The text on some of the food packaging appears slightly blurred or distorted in certain areas, particularly around the edges where there are black marks visible.
| 6.324555
|
||
130,638
|
Remove red candle
| 8
| 2
|
The edited image successfully removed the red candle from both images. The only issue is that there are some small remnants of color artifacts left near where the candle was originally placed in the original photo.
| 2
| 9
|
The image is quite effective in terms of technical quality and artifacts. The AI-generated objects are visually appealing, with a high level of detail that closely resembles real-life images. There are minimal noticeable distortions or unusual body parts/proportions; however, the overall context may seem slightly unnatural due to its artificial nature.
| 4.24264
|
||
130,638
|
Remove snack cracker box
| 4
| 6
|
The editing instruction was to remove the snack cracker box. The original image had a visible red and white Sainsbury's Savoury Snack Cracker package on top of the cheese wheel in front center-right position. In the edited version provided, this packaging has been removed from that central location but still present at bottom right with other cheeses.
| 4
| 9
|
The image is highly effective in creating a realistic and visually appealing scene of various cheeses, crackers, and wine. The textures are well-defined, the colors vibrant yet naturalistic, and there's a strong sense of depth due to careful lighting that highlights each object without introducing any noticeable artifacts or distortions.
| 6
|
||
130,638
|
Remove glass of whiskey
| 0
| 10
|
The first image shows an AI-generated picture of a table with snacks and drinks. The edited version appears to be the same as the original in terms of content but has been slightly overedited or adjusted for clarity (the cheese container is clearer). Thus, there's no change from the editing instruction ('Remove glass of whiskey'), so both scores are at their extremes: score1 = 0 because nothing was removed; score2 = 10 due to minimal yet effective edit.
| 0
| 8
|
The AI-generated image has a few minor distortions and unnatural elements, particularly noticeable in the background where some of the items appear slightly out-of-focus or unrealistic. However, for the most part, the technical quality is high with no significant artifacts affecting clarity.
| 0
|
||
92,850
|
Remove large rock
| 4
| 7
|
The edited image appears to have partially removed the large rock from one side of the gorge. However, some remnants or imperfections may still be visible around its former location on both sides of the canyon walls and at the base where it would naturally fall into the river below.
| 4
| 8
|
The image has a very high level of detail and realism, with well-rendered textures on the rock formations. The greenery around the water adds to its natural feel. However, there are some small inconsistencies in lighting that suggest AI-generated artifacts.
| 5.656854
|
||
92,850
|
Remove tree
| 9
| 1
|
The image successfully shows a waterfall flowing through a narrow canyon with moss-covered rocks and cliffs. The editing instruction was to remove trees from the scene, which has been done effectively as there are no visible tree trunks or branches in this edited version.
| 1
| 9
|
The image displays an impressive natural scene with a waterfall flowing through a narrow canyon. The AI-generated elements, such as the moss and vegetation on the rocks, are well-integrated into the environment without noticeable distortions or unnatural proportions. There is minimal evidence of artifacts that could indicate machine generation.
| 3
|
||
92,850
|
Remove waterfall
| 9
| 1
|
The original image shows a waterfall flowing through the rocks. The edited version appears to be almost identical with only slight variations in color and texture around the mossy areas.
| 1
| 10
|
The image is highly detailed and showcases natural elements with realistic textures, colors, and lighting. There are no noticeable artifacts or distortions that would indicate AI generation.
| 3.162278
|
||
92,850
|
Remove bushes
| 10
| 8
|
The original image features a waterfall flowing through a narrow canyon with vegetation and rock formations. The edited version successfully removes the bushes without altering other elements of the scene significantly, achieving high success in removing unwanted objects while maintaining overall composition.
| 8
| 9
|
The image is highly detailed and captures the natural beauty of a narrow canyon with flowing water. The textures, colors, and lighting are realistic but there appear to be some minor inconsistencies in how certain elements interact within their environment that hint at AI generation artifacts such as slight blurring or unnatural smoothness around small details.
| 8.485281
|
||
92,850
|
Remove fern
| 9
| 1
|
The editing was successful in removing the fern from the scene. The original image had a noticeable amount of greenery that seemed out of place compared to the rest of the environment. In contrast, the edited version has this greenery removed cleanly and seamlessly integrated with the overall aesthetic.
| 1
| 10
|
The image is of a waterfall flowing through an impressive rock formation, with lush greenery surrounding it. The AI-generated elements appear seamless and natural without any noticeable distortions or artifacts.
| 3.162278
|
||
109,027
|
Remove the girl
| 0
| 10
|
The edited image has successfully removed the girl as per the instruction. The scene in this version accurately follows the original without any visible human figure present.
| 0
| 9
|
The image has very few visible distortions or unusual body parts and proportions. The scene is artistically rendered with a beautiful, natural landscape under the night sky. There are some minor issues such as slightly unnatural object shapes like the tree branches in the foreground but overall it appears to be of high quality.
| 0
|
||
109,027
|
Remove bus stop sign
| 10
| 8
|
The edited image successfully removes the bus stop sign as instructed. The scene in the edited version closely follows the original with minimal changes and retains its overall aesthetic appeal.
| 8
| 8
|
The image showcases a beautiful and dreamy scene with high-quality rendering, vivid colors, and detailed elements. The girl's pose is graceful and natural-looking. However, there are minor distortions in the reflection of trees on the water surface which indicate some AI artifacts.
| 8
|
||
109,027
|
Remove trees
| 5
| 7
|
The edited image successfully removed the trees as per the instruction. However, there was a significant amount of overediting visible in some areas where elements were replaced or altered unnaturally.
| 5
| 9
|
The image is highly detailed and has a beautiful color palette, creating an immersive scene. There are no significant artifacts or distortions visible in the image.
| 6.708204
|
||
109,027
|
Remove the dress
| 10
| 9
|
The edited image successfully follows the editing instruction to remove the dress from the character without altering other elements significantly. The overall scene remains coherent and visually appealing with minimal overediting.
| 9
| 9
|
The image is highly effective with minimal artifacts. The details in the character, trees, and background blend seamlessly without noticeable distortions or unnatural object shapes. However, there are minor areas where AI-generated elements can be identified by slight inconsistencies in texture rendering.
| 9
|
||
109,027
|
Remove braided hair
| 8
| 9
|
The edited image successfully removes the braided hair from the character without significantly altering other elements of the scene. The score for editing success is high because it effectively addresses the instruction but still preserves most details and context in the background. The degree of overediting also receives a high score as there are no drastic changes to the original composition.
| 8
| 8
|
The image is visually stunning and artistically rendered, with a beautiful sky transition from blue to orange. The AI-generated trees and foliage appear naturalistic but not flawless in detail compared to hand-painted artistry. There are some minor distortions near the signpost where elements slightly overlap or have unusual proportions. However, these artifacts do not detract significantly from the overall quality.
| 8
|
||
138,265
|
Remove ladder
| 7
| 6
|
The editing instruction was to remove the ladder. In the edited image, the ladder is indeed removed or concealed effectively without altering other elements significantly. This indicates a successful execution of the editing instruction with minimal disruption in surrounding details and layout.
| 6
| 6
|
The image shows a well-constructed model of an urban area, but there are noticeable artifacts and distortions. The structures seem to be slightly misaligned or overlapping in some places which creates visual inconsistencies. Additionally, the textures on certain parts appear unnatural compared to others. However, overall it maintains a cohesive representation typical for such models.
| 6
|
||
138,265
|
Remove blueprints
| 0
| 10
|
The editing instruction was to remove blueprints. In the edited image, all traces of blueprints have been removed completely and cleanly from both images. The scene in the edited image follows the instruction perfectly as no blueprint is present.
| 0
| 7
|
The image has some noticeable artifacts, such as slight distortions in the gradient of colors and a lack of sharpness around edges. The background shows minor inconsistencies which might indicate AI-generated elements. However, overall, it is not significantly affected by these issues.
| 0
|
||
138,265
|
Remove cardboard model
| 4
| 5
|
The edited image successfully removed the cardboard model by replacing it with a plain beige board. This aligns well with the editing instruction of removing the cardboard object from the scene. However, there was some over-editing evident in the replacement process as evidenced by noticeable artifacts and inconsistencies around where the original elements were located.
| 4
| 8
|
The image has a relatively high level of realism, with well-defined edges and textures. However, there are some subtle artifacts present in the lighting reflections on surfaces like the wall and ladder, indicating it is an AI-generated image.
| 5.656854
|
||
138,265
|
Remove black bags
| 8
| 2
|
The edited image effectively removes the black bags as instructed. The scene is recognizable and follows the editing instruction text perfectly (score: 10). However, there are minor discrepancies in lighting or color that slightly overedit parts of the model, reducing it from a perfect match to an almost unrecognizable version.
| 2
| 7
|
The image shows a model of an urban environment made from cardboard, with visible distortions and irregularities in the shapes. The overall composition is somewhat cluttered but functional for representing buildings and streets. There are no apparent unusual body parts or proportions as mentioned earlier.
| 3.741657
|
||
138,265
|
Remove ceiling beam
| 6
| 8
|
The edited image successfully removes the ceiling beam as instructed. The scene in the edited image follows the editing instruction text perfectly (score: 10). However, there is some overediting where elements such as details on the model are not entirely accurate or proportionally altered which might affect the overall look of the model slightly less naturally than desired (score: 7)
| 6
| 7
|
The image shows a well-constructed model of an urban area, with clear attention to detail and proportion. The use of cardboard for buildings adds texture and depth while maintaining structural integrity. However, there are minor distortions in the alignment between different parts of the scene which may suggest slight inconsistencies or imperfections during construction.
| 6.480741
|
README.md exists but content is empty.
- Downloads last month
- 6