Hierarchical Representations for Spatio-Temporal Visual Attention Modeling and Understanding
This work addresses the challenge of understanding visual attention in videos, which is incremental as it builds on existing attention modeling approaches.
The thesis tackled the problem of modeling visual attention in video sequences by proposing two computational models: a generative probabilistic model for context-aware attention and a deep network architecture for spatio-temporal attention estimation, resulting in improved attention modeling in the temporal domain.
This PhD. Thesis concerns the study and development of hierarchical representations for spatio-temporal visual attention modeling and understanding in video sequences. More specifically, we propose two computational models for visual attention. First, we present a generative probabilistic model for context-aware visual attention modeling and understanding. Secondly, we develop a deep network architecture for visual attention modeling, which first estimates top-down spatio-temporal visual attention, and ultimately serves for modeling attention in the temporal domain.