arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.04612cs.CV
收藏

超越最大间隔:Few-Shot目标检测的类间隔均衡

Beyond Max-Margin: Class Margin Equilibrium for Few-shot Object Detection

University of Chinese Academy of Sciences(中国科学院大学);Xiamen University(厦门大学);Peng Cheng Laboratory(鹏城实验室)

Bohao Li, Boyu Yang, Chang Liu, Feng Liu, Rongrong Ji, Qixiang Ye

摘要:

Few-Shot目标检测通过使用在一组基类对象上学习的特征表示来表示新的类对象,取得了实质性的进展。然而,新的类分类和表示之间的一个隐含的矛盾不幸被忽视。一方面,为了实现准确的新类分类,两个基类的分布必须彼此远离(最大间隔)。另一方面,为了精确地表示新类,基类的分布应该彼此接近,以减少新类的类内距离(最小边缘)。在本文中,我们提出了一个类边缘均衡(CME)的方法,其目的是在一个系统的方式来优化特征空间划分和新的类重建。CME首先通过使用全连接层来解耦定位特征,将Few-Shot检测问题转换为Few-Shot分类问题。然后,CME通过在特征学习期间引入简单而有效的类余量损失来为新类保留足够的余量空间。最后,CME追求利润平衡,以对抗性的最小-最大的方式干扰新的类实例的功能。在Pascal VOC和MS-COCO数据集上的实验表明,CME在两个基线检测器上有显著的改进(平均达到$3\sim 5\%$),达到了最先进的性能。代码可在https://github.com/Bohao-Lee/CME上获得。

英文摘要:

Few-shot object detection has made substantial progressby representing novel class objects using the feature representation learned upon a set of base class objects. However,an implicit contradiction between novel class classification and representation is unfortunately ignored. On the one hand, to achieve accurate novel class classification, the distributions of either two base classes must be far away fromeach other (max-margin). On the other hand, to precisely represent novel classes, the distributions of base classes should be close to each other to reduce the intra-class distance of novel classes (min-margin). In this paper, we propose a class margin equilibrium (CME) approach, with the aim to optimize both feature space partition and novel class reconstruction in a systematic way. CME first converts the few-shot detection problem to the few-shot classification problem by using a fully connected layer to decouple localization features. CME then reserves adequate margin space for novel classes by introducing simple-yet-effective class margin loss during feature learning. Finally, CME pursues margin equilibrium by disturbing the features of novel class instances in an adversarial min-max fashion. Experiments on Pascal VOC and MS-COCO datasets show that CME significantly improves upon two baseline detectors (up to $3\sim 5\%$ in average), achieving state-of-the-art performance. Code is available at https://github.com/Bohao-Lee/CME .

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。

↑