Duplicate from fffiloni/scene-edit-detection

Co-authored-by: Sylvain Filoni <fffiloni@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,31 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,13 @@
+---
+title: Scene Edit Detection
+emoji: ✂️ 🎞
+colorFrom: pink
+colorTo: blue
+sdk: gradio
+sdk_version: 3.4
+app_file: app.py
+pinned: false
+duplicated_from: fffiloni/scene-edit-detection
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,154 @@
+# Dependencies, see also requirement.txt ;)
+import gradio as gr
+import cv2 
+import numpy as np
+import os
+
+from scenedetect import open_video, SceneManager
+from scenedetect.detectors import ContentDetector
+
+from moviepy.video.io.ffmpeg_tools import ffmpeg_extract_subclip
+
+# —————————————————————————————————————————————————
+
+title = "Scene Edit Detection"
+description = "<p style='text-align: center'>Gradio demo of PySceneDetect. <br />Automatically find every shots in a video sequence</p><p style='text-align: center'> 1. gives you timecode in/out for each shot. 2. saves each shot as a splitted mp4 video chunk for you to download. 3. diplays a thumbnail for each shot as a gallery output.<br /> <img id='visitor-badge' alt='visitor badge' src='https://visitor-badge.glitch.me/badge?page_id=gradio-blocks.scene-edit-detection' style='display: inline-block'/></b></p>"
+article = "<p style='text-align: center'><a href='http://scenedetect.com/en/latest/' target='_blank'>PySceneDetect website</a> | <a href='https://github.com/Breakthrough/PySceneDetect' target='_blank'>Github Repo</a></p>"
+
+# —————————————————————————————————————————————————
+
+# SET INPUTS
+video_input = gr.Video(source="upload", format="mp4", label="Video Sequence", mirror_webcam=False)
+threshold = gr.Slider(label="Threshold pixel comparison: if exceeded, triggers a scene cut. Default: 27.0", minimum=15.0, maximum=40.0, value=27.0)
+
+# —————————————————————————————————————————————————
+
+def convert_to_tuple(list):
+    return tuple(list);
+
+
+def find_scenes(video_path, threshold):
+    # file name without extension
+    filename = os.path.splitext(os.path.basename(video_path))[0]
+    # Open our video, create a scene manager, and add a detector.
+    video = open_video(video_path)
+    scene_manager = SceneManager()
+    scene_manager.add_detector(
+        ContentDetector(threshold=threshold))
+    
+    # Start detection 
+    scene_manager.detect_scenes(video, show_progress=True)
+    scene_list = scene_manager.get_scene_list()
+    
+    # Push the list of scenes into data_outputs
+    data_outputs.append(scene_list)
+    gradio_components_outputs.append("json")
+    #print(scene_list)
+    
+    timecodes = []
+    timecodes.append({"title": filename + ".mp4", "fps": scene_list[0][0].get_framerate()})
+    
+    shots = []
+    stills = []
+    
+    # For each shot found, set entry and exit points as seconds from frame number
+    # Then split video into chunks and store them into shots List
+    # Then extract first frame of each shot as thumbnail for the gallery
+    for i, shot in enumerate(scene_list):
+        
+        # STEP 1
+        # Get timecode in seconds
+        framerate = shot[0].get_framerate()
+        shot_in = shot[0].get_frames() / framerate
+        shot_out = shot[1].get_frames() / framerate
+        
+        tc_in = shot[0].get_timecode()
+        tc_out = shot[1].get_timecode()
+        
+        frame_in = shot[0].get_frames()
+        frame_out = shot[1].get_frames()
+        
+        timecode = {"tc_in": tc_in, "tc_out": tc_out, "frame_in": frame_in, "frame_out": frame_out}
+        timecodes.append(timecode)
+        
+        # Set name template for each shot
+        target_name = "shot_" + str(i+1) + "_" + str(filename) + ".mp4"
+        
+        # Split chunk
+        ffmpeg_extract_subclip(video_path, shot_in, shot_out, targetname=target_name)
+        
+        # Push chunk into shots List
+        shots.append(target_name)
+        
+        # Push each chunk into data_outputs
+        data_outputs.append(target_name)
+        gradio_components_outputs.append("video")
+        
+        # —————————————————————————————————————————————————
+        
+        # STEP 2
+        # extract first frame of each shot with cv2
+        vid = cv2.VideoCapture(video_path)
+        fps = vid.get(cv2.CAP_PROP_FPS)
+        print('frames per second =',fps)
+        
+        frame_id = shot[0].get_frames() # value from scene_list from step 1
+
+        vid.set(cv2.CAP_PROP_POS_FRAMES, frame_id)
+        ret, frame = vid.read()
+
+        # Save frame as PNG file
+        img = str(frame_id) + '_screenshot.png'
+        cv2.imwrite(img,frame)
+        
+        # Push image into stills List
+        stills.append((img, 'shot ' + str(i+1)))
+    
+    # Push the list of video shots into data_outputs for Gradio file component
+    data_outputs.append(shots)
+    gradio_components_outputs.append("file")
+        
+    # Push the list of still images into data_outputs
+    data_outputs.append(stills)
+    gradio_components_outputs.append("gallery")
+    
+    # This would have been used as gradio outputs, 
+    # if we could set number of outputs after the interface launch
+    # That's not (yet ?) possible
+    results = convert_to_tuple(data_outputs)    
+    print(results)
+ 
+    # return List of shots as JSON, List of video chunks, List of still images  
+    # *
+    # Would be nice to be able to return my results tuple as outputs, 
+    # while number of chunks found is not fixed:
+    # return results 
+    return timecodes, shots, stills
+
+# —————————————————————————————————————————————————
+
+# SET DATA AND COMPONENTS OUTPUTS
+
+# This would be filled like this:
+# data_outputs = [ [List from detection], "video_chunk_n0.mp4", "video_chunk_n1.mp4", ... , "video_chunk_n.mp4", [List of video filepath to download], [List of still images from each shot found] ]
+data_outputs = []
+
+# This would be filled like this:
+# gradio_components_outputs = [ "json", "video", "video", ... , "video", "file", "gallery" ]
+gradio_components_outputs = []
+
+
+#SET OUTPUTS
+
+# This would be nice if number of outputs could be set after Interface Launch: 
+# because we do not know how many shots will be detected         
+# gradio_components_outputs = [ "json", "video", "video", ... , "video", "file", "gallery" ]
+# outputs = gradio_components_outputs
+
+# ANOTHER SOLUTION WOULD BE USING A (FUTURE ?) "VIDEO GALLERY" GRADIO COMPONENT FROM LIST :)
+
+outputs = [gr.JSON(label="Shots detected"), gr.File(label="Downloadable Shots"), gr.Gallery(label="Still Images from each shot").style(grid=3)]
+
+# —————————————————————————————————————————————————
+print('Hello Sylvain')
+gr.Interface(fn=find_scenes, inputs=[video_input, threshold], outputs=outputs, title=title, description=description, article=article).launch()

requirements.txt

@@ -0,0 +1,4 @@
+opencv-python>=4.4.0
+scenedetect<0.7
+moviepy
+imageio-ffmpeg