JPG Image Extraction from Documents
Budget: $30 – $250 SGD
Job Description: Math Quiz Image Extraction from Scanned Documents
We are seeking a skilled Python developer to create a robust script for automatically extracting individual math quiz questions from scanned documents in JPG or PDF format.
Key Responsibilities:
1. Input Handling:
- Accept both JPG and PDF files as input.
- Read files within a specified folder.
2. Question Extraction:
- Detect and isolate individual question image boundaries within the scanned documents.
- Extract each question as a separate image, maintaining the aspect ratio and quality.
3. Output Generation:
- Save extracted question images as individual image files in PNG or JPG format in a designed folder
4. Error Handling and Logging:
- Implement robust error handling.
- Provide detailed logging for troubleshooting.
5. Performance Optimization:
- Ensure the script can efficiently process all the images in the designated folder
6. Documentation:
- Provide clear documentation on how to use the script.
- Include requirements and setup instructions.
Deliverables:
- Python script(s) for image extraction.
- Requirements file listing all necessary dependencies.
- README file with usage instructions and examples.
- Sample output demonstrating the script's capabilities.
Requirements:
- Experience in image extraction and conversion.
- Proficiency in maintaining image quality.
- Knowledge of image formats, particularly JPEG.
- Ability to handle JPG and PDF files.
- Understanding of specific image requirements for web usage.
Enhancement Tools:
- OCR methods or LLM (e.g., GPT-4 or GPT-3.5).
- Segment Anything Model (SAM2) from Meta for image segmentation.
Attached Sample:
- A sample question JPG with 1 math questions, demonstrating the expected output and corresponding filenames (e.g., 72-5.jpg and 72-A1.jpg). More images in the files section (10 pages or so, those are the scanned original docs to extract from)
If you can proficiently extract images from JPG and PDF files, ensuring high-quality outputs suitable for online use, we would love to hear from you. I shall be inclined to select the freelancer who is able to provide a screencast or demo if you can do it :) thanks
We are seeking a skilled Python developer to create a robust script for automatically extracting individual math quiz questions from scanned documents in JPG or PDF format.
Key Responsibilities:
1. Input Handling:
- Accept both JPG and PDF files as input.
- Read files within a specified folder.
2. Question Extraction:
- Detect and isolate individual question image boundaries within the scanned documents.
- Extract each question as a separate image, maintaining the aspect ratio and quality.
3. Output Generation:
- Save extracted question images as individual image files in PNG or JPG format in a designed folder
4. Error Handling and Logging:
- Implement robust error handling.
- Provide detailed logging for troubleshooting.
5. Performance Optimization:
- Ensure the script can efficiently process all the images in the designated folder
6. Documentation:
- Provide clear documentation on how to use the script.
- Include requirements and setup instructions.
Deliverables:
- Python script(s) for image extraction.
- Requirements file listing all necessary dependencies.
- README file with usage instructions and examples.
- Sample output demonstrating the script's capabilities.
Requirements:
- Experience in image extraction and conversion.
- Proficiency in maintaining image quality.
- Knowledge of image formats, particularly JPEG.
- Ability to handle JPG and PDF files.
- Understanding of specific image requirements for web usage.
Enhancement Tools:
- OCR methods or LLM (e.g., GPT-4 or GPT-3.5).
- Segment Anything Model (SAM2) from Meta for image segmentation.
Attached Sample:
- A sample question JPG with 1 math questions, demonstrating the expected output and corresponding filenames (e.g., 72-5.jpg and 72-A1.jpg). More images in the files section (10 pages or so, those are the scanned original docs to extract from)
If you can proficiently extract images from JPG and PDF files, ensuring high-quality outputs suitable for online use, we would love to hear from you. I shall be inclined to select the freelancer who is able to provide a screencast or demo if you can do it :) thanks