image imagewidth (px) 268 2.06k | image_file stringlengths 41 88 | prompt stringlengths 2.5k 2.59k | visual_info_type stringclasses 2 values | ocr_text_type stringclasses 3 values |
|---|---|---|---|---|
5M_used_parts/img/pierwszy_DSC0011_93.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
24467 F1=300N
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | primary | |
5M_used_parts/img/3ad517e44f07ba70d97ef53136fe_1_1.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
GM 22795012 004-18-Z-00-00
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| secondary | primary | |
golden_set_with_ocr/98f6d81212795e07c6968b3de742de3d5af96dda8ea0e5ce37ca740852652fbc.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
FCL694958
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| secondary | primary | |
golden_set_with_ocr/ad0c5e69603e03d2654ad34dded7de0fc5a7777b9f6db110aaa8360de5034b36.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
1 St/Pc 920 0 250 203 002-EAF 955 DURATERM 016 FIAT OPEL SUZUKI 18€
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| secondary | primary | |
golden_set_with_ocr/c7ae17d195c38538c6a585cd4873d81a88d4fc564d62b8e24e8d6ae6be4be16e.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
076 906 051B VW AG 46/11 TN3
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| secondary | primary | |
golden_set_with_ocr/7db9829ef27606c344ce5a45f1e79c4edbbeebe9842efd7f3a20366219fd5ed3.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
RAVENOL
ATF
Art. 1212100-010
Automatik GetriebeöI
Dexron III H
SYNTHETIC
MADE IN GERMANY
10L
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | primary | |
5M_used_parts/img/b0602705472ba264a23b1d8e3586_1.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
corsa d. 7000
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | secondary | |
golden_set_with_ocr/1644c0135b98fc906620fe2b8356acd0478522e60d3aa784be9db73073b20b14.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
Castrol TRANSMAX ATF DX III MULTIVEHICLE AUTOMATIC TRANSMISSION FLUID 1L FindGearboxOil.com
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | primary | |
5M_used_parts/img/b46172824d70a68612e95a1a4a22_1.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
A0008200404
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | secondary | |
5M_used_parts/img/pierwszy_DSC0355_35.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
001182
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | secondary | |
5M_used_parts/img/pierwszyDSC_0354_979.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
nan
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | no_text | |
5M_used_parts/img/69a584484a91bb365208b665c202_1_1.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
nan
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | no_text | |
5M_used_parts/img/pierwszyDSC_0552_552.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
nan
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | no_text | |
5M_used_parts/img/pierwszyDSC_0017_1186.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
nan
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | no_text | |
5M_used_parts/img/02c67e0340bba86821bb0cfa0c92_1.jpg |
# TASK
You are given an image and the OCR text extracted from it. Follow the steps below to produce a structured data response.
# OCR TEXT
```
nan
```
# STEPS
1. Analyze the image and the provided OCR text.
2. Generate a single JSON object that strictly conforms to the schema below.
3. Include all required fields from the schema.
4. Adhere to the specified types and enumerations exactly.
5. Return only the JSON object (no extra explanations or text).
# JSON SCHEMA
{
"$schema": "http://json-schema.org/draft-07/schema#",
"type": "object",
"title": "Image Analysis Schema",
"description": "Schema for extracting structured data from an image, including justifications for value assignments.",
"properties": {
"image_description": {
"type": "string",
"description": "A short and concise description of the image."
},
"visual_information_type": {
"type": "string",
"enum": ["primary", "secondary", "incomprehensible"],
"description": "Classify the visual information in the image based on its criticality for product identification.",
"examples": [
"primary: The image shows an object like automotive, mechanical or furniture object that is clearly visible with high-quality details.",
"secondary: The image has rich structured text like documents, barcodes, alphanumerics or a partially visible object.",
"incomprehensible: The image is too unclear to interpret anything."
]
},
"reason_for_visual_information_type": {
"type": "string",
"description": "Explain why the selected visual information type applies to this image."
},
"ocr_text_type": {
"type": "string",
"enum": ["primary", "secondary", "not_relevant"],
"description": "Classify the OCR text based on its importance for product identification.",
"examples": [
"primary: The OCR text includes serial numbers, product codes, bar codes, VINs, or product names.",
"secondary: The OCR text includes websites, brand names, telephone numbers, addresses, etc.",
"not_relevant: The OCR text is irrelevant to product identification."
]
},
"reason_for_ocr_text_type": {
"type": "string",
"description": "Explain why the selected OCR text type applies to this text."
},
},
"required": [
"image_description",
"visual_information_type",
"reason_for_visual_information_type",
"ocr_text_type",
"reason_for_ocr_text_type",
]
}
| primary | no_text |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.