Social Applications
The analysis of visual information, such as images and video, finds diverse social, behavioral and societal applications. For example, image analysis to find frequent appearances of the same groups of people in photos can help establish social networks, while facial expression analysis from images/video can be beneficial for assessing and tracking mood, stress, etc.
In our collaborative research with Boston University’s School of Social Work, we have studied image analysis approaches to help with the diagnosis and treatment of the hoarding disorder.
- Image Analysis for Hoarding Assessment: Room clutter classification from images using ResNet-18.
- Improved Indoor Clutter Classification from Images: Improved classification of indoor clutter related to the hoarding disorder from images using the Visual Transformer.
- Datasets of images with indoor clutter related to hoarding annotated by health professionals according to Clutter Image Rating (CIR) scale: HINDER-2018 and HINDER-2025.