Applied Computer Vision Services
A collection of image systems built around concrete operations.
Built and integrated computer vision services for image restoration, upscaling, colorization, captioning, OCR, face detection, background removal, and content moderation.
These were applied systems rather than one single research project. Each one had to answer a practical question about what an image service should return and how another application could use it.
Document and image outputs
The OCR work returned both recognized text and its locations in the image. Other services returned a restored image, a caption, detected faces, a removed background, or a moderation result.
From model to service
The work connected model inference to web APIs and small applications so the result could be called by software instead of remaining inside a training notebook.
selected tools
- Python
- PyTorch
- OpenCV
- MMOCR
- FastAPI