4.8 Article

Learning Deep Binary Descriptors via Bitwise Interaction Mining

出版社

IEEE COMPUTER SOC
DOI: 10.1109/TPAMI.2022.3161600

关键词

Reliability; Reinforcement learning; Costs; Binary codes; Training; Task analysis; Semantics; Binary descriptors; unsupervised learning; bitwise interaction; reinforcement learning; differentiable search; graph convolutional networks

向作者/读者索取更多资源

In this paper, we propose a GraphBit method for learning unsupervised deep binary descriptors to efficiently represent images. The method reduces the uncertainty of binary codes by maximizing the mutual information with input and related bits, allowing reliable binarization of ambiguous bits. Additionally, a differentiable search method called GraphBit+ is introduced to mine bitwise interaction in continuous space, reducing the computational cost of reinforcement learning. To address the issue of inaccurate instructions from fixed bitwise interaction, the unsupervised binary descriptor learning method D-GraphBit is proposed, which utilizes a graph convolutional network to reason the optimal bitwise interaction for each input sample.
In this paper, we propose a GraphBit method to learn unsupervised deep binary descriptors for efficient image representation. Conventional binary representation learning methods directly quantize each element according to the threshold without considering the quantization ambiguousness. The elements near the boundary dubbed as ambiguous bits fail to collect effective information for reliable binarization and are sensitive to noise that causes reversed bits. We argue that there are implicit inner relationships among bits in binary descriptors called bitwise interaction, where the related bits can provide extra instruction as prior knowledge for ambiguousness reduction. Specifically, we design a deep reinforcement learning model to learn the structure of the graph for bitwise interaction mining, and the uncertainty of binary codes is reduced by maximizing the mutual information with input and related bits. Consequently, the ambiguous bits receive additional instruction from the graph for reliable binarization. Moreover, we further present a differentiable search method (GraphBit+) that mines the bitwise interaction in continuous space, so that the heavy search cost caused by the training difficulties in reinforcement learning is significantly reduced. Since the GraphBit and GraphBit+ methods learn fixed bitwise interaction which is suboptimal for various input, the inaccurate instruction from the fixed bitwise interaction cannot effectively decrease the ambiguousness of binary descriptors. To address this, we further propose the unsupervised binary descriptor learning method via dynamic bitwise interaction mining (D-GraphBit), where a graph convolutional network called GraphMiner reasons the optimal bitwise interaction for each input sample. Extensive experimental results on the CIFAR-10, NUS-WIDE, ImageNet-100, Brown and HPatches datasets demonstrate the efficiency and effectiveness of the proposed GraphBit, GraphBit+ and D-GraphBit.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.8
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据