Ability to remove sensitive information and background from images

[zhiltsov-max]

When publishing a dataset and when training a model (to avoid bias) it could be useful to remove sensitive information like car plate numbers, names and other personal info on medical images, faces etc. Another variant - is removal of background (noise) from images for anomaly detection tasks. One more - removal of reflections on the image.

[zhiltsov-max]

OpenCV G-API now (in 5.0) provides the option to blur faces efficiently, it can be useful for this feature

• infer https://github.com/opencv/opencv/blob/master/modules/gapi/misc/python/test/test_gapi_infer.py#L35
• parseSSD
  • https://docs.opencv.org/4.5.3/d4/d1c/namespacecv_1_1gapi.html#ae6693e4d29816df39cc5547f400c88c7
  • https://github.com/opencv/opencv/blob/master/modules/gapi/misc/python/test/test_gapi_infer.py#L312
• Custom kernel (ToMosaic)
  • https://github.com/opencv/opencv/blob/master/modules/gapi/misc/python/test/test_gapi_sample_pipelines.py#L26-L41
  • https://github.com/opencv/opencv/blob/master/modules/gapi/misc/python/test/test_gapi_sample_pipelines.py#L223-L227
• Rendering: https://github.com/opencv/opencv/blob/master/modules/gapi/misc/python/test/test_gapi_render.py#L17