I’ve added my university honours project to this website. It is called Modern to Historical Image Feature Matching. The project can be briefly described as an attempt at understanding why modern computer vision techniques aren’t successful at matching historical photos taken on film with modern digital photos. The project includes a paper (linked above), source code and a data set. The project was written in C++, using the OpenCV computer vision libraryAll the appropriate links can be found in the programming section of this website. The following is the abstract of the project paper:
Matching images with each other using their unique features has many applications. While there currently exists many robust image
feature matching systems, we propose that they are currently only accurate when matching modern photos with each other. Historical
photos, traditionally taken on ﬁlm, when converted to digital format are not likely to successfully match with modern equivalents. Photos of famous landmarks from various dates were gathered to verify this hypothesis. Images were matched using four standard feature matching systems. These systems matched either image corners or image blobs with each other. Match successes were recorded for each image comparison, and analyzed in tables. The results showed that the feature systems accurately matched modern photos with each other. However, when matching historical photos to modern photos, results were poor. A novel approach to feature matching using image regions was proposed. The expectation was that using regions instead of corners or blobs would result in higher accuracy when matching. Initial results were mixed, and no concrete solution was found.