Deep Specialized Network for Illuminant Estimation | |
Wu Shi; Chen Change Loy; Xiaoou Tang | |
2016 | |
会议名称 | ECCV2016 |
会议地点 | 荷兰阿姆斯特丹 |
英文摘要 | Illuminant estimation to achieve color constancy is an ill- posed problem. Searching the large hypothesis space for an accurate illu- minant estimation is hard due to the ambiguities of unknown re ections and local patch appearances. In this work, we propose a novel Deep Specialized Network (DS-Net) that is adaptive to diverse local regions for estimating robust local illuminants. This is achieved through a new convolutional network architecture with two interacting sub-networks, i.e. an hypotheses network (HypNet) and a selection network (SelNet). In particular, HypNet generates multiple illuminant hypotheses that inher- ently capture di erent modes of illuminants with its unique two-branch structure. SelNet then adaptively picks for con dent estimations from these plausible hypotheses. Extensive experiments on the two largest color constancy benchmark datasets show that the proposed `hypoth- esis selection' approach is e ective to overcome erroneous estimation. Through the synergy of HypNet and SelNet, our approach outperforms state-of-the-art methods such as [1{3]. |
收录类别 | EI |
语种 | 英语 |
内容类型 | 会议论文 |
源URL | [http://ir.siat.ac.cn:8080/handle/172644/10026] |
专题 | 深圳先进技术研究院_集成所 |
作者单位 | 2016 |
推荐引用方式 GB/T 7714 | Wu Shi,Chen Change Loy,Xiaoou Tang. Deep Specialized Network for Illuminant Estimation[C]. 见:ECCV2016. 荷兰阿姆斯特丹. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论