Python OCR & IP Camera Integration
Budget: ₹37,500 – ₹75,000 INR
We need a professional with:
Strong expertise in Python
Proven experience with AI-based OCR (must have worked on similar projects earlier)
Knowledge of camera integration
Preference: Based in Mumbai or Pune, available for regular on-site visits for installation and deployment.
For more details, message me directly Wats.Ap: +91 88.289.599.43.
Deliverables
1. Fully commented Python source code that opens the RTSP stream, stabilises frames if needed, detects regions of interest, performs OCR, and returns the recognised text in UTF-8.
2. A short README (markdown is fine) explaining setup, configuration, and how to extend or retrain the model.
3. A quick video or screenshot set demonstrating the script in action on both printed signage and handwritten notes held up to the camera.
Acceptance criteria
• ≥ 90 % character-level accuracy on my sample test set (mixed print/handwriting).
• No dropped frames for streams up to 1080p@30 fps.
• Clean exit and automatic reconnection if the stream temporarily goes down.
If you have already tackled similar real-time OCR or vision projects, share a concise example when you respond so I can gauge fit quickly.
Strong expertise in Python
Proven experience with AI-based OCR (must have worked on similar projects earlier)
Knowledge of camera integration
Preference: Based in Mumbai or Pune, available for regular on-site visits for installation and deployment.
For more details, message me directly Wats.Ap: +91 88.289.599.43.
Deliverables
1. Fully commented Python source code that opens the RTSP stream, stabilises frames if needed, detects regions of interest, performs OCR, and returns the recognised text in UTF-8.
2. A short README (markdown is fine) explaining setup, configuration, and how to extend or retrain the model.
3. A quick video or screenshot set demonstrating the script in action on both printed signage and handwritten notes held up to the camera.
Acceptance criteria
• ≥ 90 % character-level accuracy on my sample test set (mixed print/handwriting).
• No dropped frames for streams up to 1080p@30 fps.
• Clean exit and automatic reconnection if the stream temporarily goes down.
If you have already tackled similar real-time OCR or vision projects, share a concise example when you respond so I can gauge fit quickly.