Sobre

## The Problem We have a facial recognition system with a dataset that's being compromised by problematic images. These bad photos are causing interference in our vector search and reducing accuracy. **Specific issues we're facing:** - Overexposed/blown-out lighting - Multiple faces in single images - Blurry or out-of-focus photos - Incorrect angles or partial faces - Poor contrast or too dark images ## What We Need A Python script that: 1. Scans through a folder of JPEG face images 2. Identifies problematic photos using quality metrics 3. Automatically moves bad images to a separate folder 4. Can be run weekly to maintain dataset cleanliness ## Important Notes - This involves sensitive customer facial data (we'll provide sample Google images for testing, but production data stays private) - Script needs to be reliable enough for automated weekly runs - Should generate a report of what was filtered and why ## Compensation Structure **Base payment:** For a functional script that correctly identifies and separates bad images **Quality bonus available:** If your script exceeds expectations with features like: - Advanced detection algorithms with high accuracy - Detailed quality reports with improvement suggestions - Efficient processing for large datasets - Clean, well-documented, maintainable code - Additional useful features we didn't think of We believe in paying well for excellent work. Deliver something exceptional, and we'll happily compensate accordingly. ## Required Skills - **Python programming** - Clean, maintainable code - **Image processing** - OpenCV, PIL/Pillow, or similar - **Quality detection algorithms** - Face detection, blur detection, exposure analysis - **Batch processing** - Efficient handling of large image folders If you're experienced in image quality assessment and enjoy solving real-world AI problems, we'd love to work with you. Please share examples of similar image processing work you've done.