I was asked by the GRI to provide my expertise in Computer Vision to help them plan their actions with digitizing and organizing their very large photo collection.

Apart from informal discussions, I spent several weeks testing and validating processes (automatic image separation, text recognition, image similarity, aligning different visual collections, etc…) culminating in a report and even a web demo showing the capabilities of Computer Vision applied to this sort of data.

The report provides a comprehensive plan of actions, detailling an effective approach to the complex task of semi-automatically organizing such a large archive. For each subprocess identified, it provides expert information about the expected difficulty to implement them, and where to prioritize work.