arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1607.01437cs.CV
收藏

自适应零件的属性识别

Attribute Recognition from Adaptive Parts

Simon Fraser University(西蒙弗雷泽大学);Zhejiang University(浙江大学);Microsoft Research Asia(微软亚洲研究院);Tongji University(同济大学)

Luwei Yang, Ligen Zhu, Yichen Wei, Shuang Liang, Ping Tan

摘要:

以前的基于零件的属性识别方法在单独的步骤中执行零件检测和属性识别。这些部件未针对属性识别进行优化,因此可能是次优的。我们提出了一种端到端的深度学习方法来克服这一限制。它从关键点生成对象部分,并相应地执行属性识别,允许自适应空间变换的部分。关键点估计和属性识别都是在多任务环境中共同学习的。在两个数据集上进行的大量实验验证了所提出的端到端方法的有效性。

英文摘要:

Previous part-based attribute recognition approaches perform part detection and attribute recognition in separate steps. The parts are not optimized for attribute recognition and therefore could be sub-optimal. We present an end-to-end deep learning approach to overcome the limitation. It generates object parts from key points and perform attribute recognition accordingly, allowing adaptive spatial transform of the parts. Both key point estimation and attribute recognition are learnt jointly in a multi-task setting. Extensive experiments on two datasets verify the efficacy of proposed end-to-end approach.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑