arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2210.05899cs.CV
收藏

Hash码性能的一个下界

A Lower Bound of Hash Codes' Performance

University of Electronic Science and Technology of China(电子科技大学);Peng Cheng Laboratory(鹏城实验室)

Xiaosu Zhu, Jingkuan Song, Yu Lei, Lianli Gao, Heng Tao Shen

摘要:

作为紧凑表示学习的一种重要方法,散列算法在有效性和效率方面取得了巨大的成功。许多启发式汉明空间度量学习目标的设计,以获得高质量的哈希码。然而,学习好的哈希码的标准的理论分析仍然在很大程度上未被利用。本文证明了散列码的类间区分性和类内紧性决定了散列码的性能下限。提升这两个特性可以提升边界并改善哈希学习。然后,我们提出了一个代理模型,充分利用上述目标,估计后验的哈希码和控制它,这导致在一个低偏差的优化。大量的实验表明了该方法的有效性。通过对一系列散列模型的测试,我们得到了所有这些模型的性能改进,平均平均精度增加了$26.5\%$,准确性增加了$20.5\%$。我们的代码可在https://github.com/VL-Group/LBHash上公开获取。

英文摘要:

As a crucial approach for compact representation learning, hashing has achieved great success in effectiveness and efficiency. Numerous heuristic Hamming space metric learning objectives are designed to obtain high-quality hash codes. Nevertheless, a theoretical analysis of criteria for learning good hash codes remains largely unexploited. In this paper, we prove that inter-class distinctiveness and intra-class compactness among hash codes determine the lower bound of hash codes' performance. Promoting these two characteristics could lift the bound and improve hash learning. We then propose a surrogate model to fully exploit the above objective by estimating the posterior of hash codes and controlling it, which results in a low-bias optimization. Extensive experiments reveal the effectiveness of the proposed method. By testing on a series of hash-models, we obtain performance improvements among all of them, with an up to $26.5\%$ increase in mean Average Precision and an up to $20.5\%$ increase in accuracy. Our code is publicly available at https://github.com/VL-Group/LBHash.

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。

↑