Bit-Scalable Deep Hashing With Regularized Similarity Learning for Image Retrieval and Person Re-Identification

By Ruimao Zhang, Liang Lin, Rui Zhang, Wangmeng Zuo, Lei Zhang. IEEE Transactions on Image Processing, 2015

492 citations recorded by the source corpus.

Extracting informative image features and learning effective approximate hashing functions are two crucial steps in image retrieval. Conventional methods often study these two steps separately, e.g., learning hash functions from a predefined hand-crafted feature space. Meanwhile, the bit lengths of output hashing codes are preset in the most previous methods, neglecting the significance level of different bits and restricting their practical flexibility. To address these issues, we propose a supervised learning framework to generate compact and bit-scalable hashing codes directly from raw images. We pose hashing learning as a problem of regularized similarity learning. In particular, we organize the training images into a batch of triplet samples, each sample containing two images with the same label and one with a different label. With these triplet samples, we maximize the margin between the matched pairs and the mismatched pairs in the Hamming space. In addition, a regularization term is introduced to enforce the adjacency consistency, i.e., images of similar appearances should have similar codes. The deep convolutional neural network is utilized to train the model in an end-to-end fashion, where discriminative image features and hash functions are simultaneously optimized. Furthermore, each bit of our hashing codes is unequally weighted, so that we can manipulate the code lengths by truncating the insignificant bits. Our framework outperforms state-of-the-arts on public benchmarks of similar image search and also achieves promising results in the application of person re-identification in surveillance. It is also shown that the generated bit-scalable hashing codes well preserve the discriminative powers with shorter code lengths.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/0267739fce0380036ee515d3310b9bc2f723b1ce

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Dacheng Tao
2. Luc Van Gool
3. Yi Yang
4. Ming-Hsuan Yang
5. Lei Zhang
6. Xiaogang Wang
7. Trevor Darrell
8. Jiashi Feng

Citations in the corpus

• A Decade Survey of Content Based Image Retrieval Using Deep Learning (2020, methodology): “CNN Features [95], [126] Neural Code [61] CNN off-the-shelf [32] 2015 Binary Autoencoder [127] Patch + Siamese [128], [129] Adjacency Consistency Triplet [64]”
• Learning Discrete Representations via Information Maximizing Self-Augmented Training (2017, background): “Regarding hashing learning, a number of studies have used deep neural networks for supervised hash learning and achieved state-of-the-art results on image and text retrievals (Xia et al., 2014; Lai et al., 2015; Zhang et al., 2015; Xu et al., 2015; Li et al., 2015).”
• Joint Learning of Single-Image and Cross-Image Representations for Person Re-identification (2016, methodology): “For the methods based on deep CNN features, feature representation and classifier can be jointly optimized for learning either SIR or CIR features.”
• Unsupervised Learning of Visual Representations Using Videos (2015, methodology): “[55] optimized the max-margin loss on triplet units to learn deep hashing function for image retrieval.”
• Unsupervised Semantic-Preserving Adversarial Hashing for Image Search (2019, methodology): “We further compare UADH with several popular supervised hashing methods, including DRSCH [70], DSCH [70], DSRH [71], and DPSH [41].”
• Deep Supervised Discrete Hashing (2017, methodology): “Similar to [8, 11, 19, 23], we use a subset of 195,834 images which are associated with the 21 most frequent concepts.”
• Person Re-identification: Past, Present and Future (2016, background): “The only relevant work we find is [158] which learns hash functions in a triplet-loss CNN network with regularizations to enforce adjacency consistency.”
• Feature Learning Based Deep Supervised Hashing with Pairwise Labels (2015, methodology): “The experimental setting in DSCH and DRSCH [Zhang et al., 2015] is different from that in Section 4.2.”