Projects
 |
Weakly Supervised Learning for Attribute Localization in Outdoor Scenes In recent years, attributes become popular because they are shared across categories and can provide rich semantics to scene representationsIn this project, we connect the attributes to the hierarchical scene model to reinforce the complementary advantages of the semantics and hierarchy.
|
 |
Hierarchical Space Tiling in Scene Modeling A typical scene category contains an enormous number of distinct scene configurations that are composed of objects and regions of varying shapes in different layouts. In this project, we first propose a representation named Hierarchical Space Tiling (HST) to quantize the huge and continuous scene configuration space.
|
 |
Human Attribute Recognition by Rich Appearance Dictionary Recognizing human attributes, fine-grained semantics such as gender, hair style, garment types, from static images is an important problem. The key challenge, however, lies in huge complexity arisen from geometric variation as well as versatile appearance types of human parts. In this project, we introduce a simple yet effective weakly-supervised part learning method.
|