Recherche partenaire principal
Human memorability is a very important aspect of life in the society. As they age people tend to forget things they have seen. This project will use human subjects, respecting the ethics of such experiments, which will see sets of images. After a while, the same participants will see another set of images indicating whether they remember the images and what regions they remember. The computational part of the experiment relies on extracting features from images, grouping them syntatically and statistically and then feeding them into a classifier. The classfier will learn the characteristics of image regions that are remembered by people and will aim to replicate them in computers. If we succeed having sufficient data sets we will use Deep Learning Networks. This project could be useful for retrieving human' memory relevant images as well as for advertising and marketing. Our expertise is in Image Processing, Machine Learning, Content-Based Image Retrieval among others.