SLIACV 2012 - The International Special Session on Structured Learning and Its Applications in Computer Vision

Structured Learning and Its Applications in Computer Vision
[Session Chairs]: Jie Liu (JieLiu-AT-nankai.edu.cn)
[Scope]: This session addresses the problem of learning from structured data in computer vision. Many important problems in computer vision involve implicitly or explicitly structured data, such as image segmentation, action recognition and object labeling. Compared with the independent identical data in the traditional machine learning task, the structured data is no more vectorial but structured: a data item is described by parts and relations between parts, where the description obeys some underlying rules. One typical example of structured data learning is action recognition, modeled as time sequence consisting of images and each image is a frame of the action video. The complex nature of structured data poses unique and unprecedented challenges to both research community and industry. This special session aims at bringing together researchers and industry practitioners in the fields of structured data learning to address particular problems and challenges in the context of object recognition and other applications.
Topics:
Images Segmentation
Action recognition
Object labeling
Medical image analysis
Behavior analysis
Expression recognition
Gesture recognition
Graphical Models (conditional random field, hidden Markov model, etc.)
Structured Support Vector Machines