報告題目：Hierarchical Learning for Large-Scale Visual Recognition
報告人：Jianping Fan
時間：2015年6月29日 9:00
地點：仙林校區(qū)行政南樓456室
主辦單位：計算機學院、軟件學院、科技處
Abstract: In this talk, I will introduce our research on large-scale visual recognition through hierarchical learning of tree classifiers. First, a visual treeis learned for organizing large number of object classes and image concepts in a coarse-to-fine fashion, which can provide a good environment for determining the inter-related learning tasks automatically in the feature space. Second, a multi-task structural learning algorithm is developed to learn the inter-related classifiers for the sibling child nodes under the same parent node. An inter-level constraint is introduced to learn more discriminative classifiers for the high-level nodes on the visual tree.
Bio: Jianping Fan got his MS degree on theory physics from Northwest University, Xi’an, China and his PhD degree on computer science from Shanghai Institute of Optics and Fine Mechanics, CAS. He is now a professor at UNC-Charlotte. His research interests include statistical machine learning, computer vision, and multimedia retrieval.
蘇公網(wǎng)安備32011302320419號 |蘇ICP備11073489號-1