Modeling local tone-mapping for raw image reconstruction-aware deep image compressors

Cameras apply a lot of processing on the raw image recorded by the sensor to enhance the brightness, contrast, and colors, and make the output image visually pleasing. The image is also finally compressed to make the file size smaller. These operations make it difficult to reverse to the raw sensor image which is necessary for several computer/machine vision tasks. Existing AI methods to invert from the camera’s compressed output image to the raw sensor image assume that only global color and tone manipulations are applied. However, most present-day consumer cameras perform local tone adjustment to further improve picture quality. The goal of this project is to devise an AI algorithm that allows accurate recovery of the raw sensor image from the camera’s locally-tone adjusted compressed output image. This project will help Samsung Electronics Canada develop an improved and more practical raw recovery AI algorithm applicable to modern DSLR and smartphone camera images.

Faculty Supervisor:

Richard Wildes

Student:

Abhijith Punnappurath

Partner:

Samsung Electronics Canada

Discipline:

Engineering - computer / electrical

Sector:

Manufacturing

University:

York University