AI - Image quality evaluation - Case Study

Before implementing a compliance project, the client wanted to evaluate if the AI model can produce the desired output. Client wanted to validate if the human validations can be replaced by AI to evaluate low quality images.

Outcome

Implemented and tuned two deep learning model
- To check if the image is blurry or not.
- To recognize faces in the image.
The output produces by our model was more than 99% successful. Out of 2000 images evaluated by the mode, client was not able to identify one issue with the output.
Provided a estimation and plan to implement the entire solution.

‍