Abstract:<p class="a-plus-plus">With the development of Internet technology, an increasing amount of data enters people's daily life, which brings great challenges when users quickly search for interesting images. The existing exact nearest neighbor retrieval methods often fail to obtain results within an acceptable retrieval time, so researchers have begun to focus on approximate nearest neighbor retrieval. Recently, the hashing-based approximate nearest neighbor retrieval method has attracted increasing attention because of its small storage space and high retrieval efficiency. At present, one of the most advanced hashing methods is to use deep neural networks, especially convolutional neural networks (CNN), to obtain image hash codes to achieve fast image retrieval. However, CNN needs a large number of images during training, so it takes a lot of time to obtain training samples. In addition, CNN cannot handle ambiguity well, and a lot of information is lost in the pooling layer; furthermore, CNN cannot learn the hierarchical structure of the image. Aiming to address these problems while making full use of data classification information to guide hash learning in a supervised form to improve retrieval efficiency, we introduce the capsule network into the hash learning and propose CapsNet-based supervised hashing (CSH) to preserve the effective information of an image as much as possible. CSH adds a hashing layer equivalent to hash mapping between the capsule network and the decoder to perform hash learning. By optimizing the objective function defined for capsule network loss, reconstruction loss and hashing quantization loss, feature representations, hashing functions and classification results can be learned from the input data at the same time. To verify the effectiveness of the method, we performed experiments on multiple datasets. The experimental results show that this method is superior to the existing hashing-based image retrieval methods and achieves satisfactory results in image classification performance.</p>

Training data ... ... 0 1 1 0 1 0 0 1 Convolution and pooling layers fc layers Hashing layer Regularized center loss Bike Car Bird Dicycle Semantic embedding loss

SitNet: Discrete Similarity Transfer Network for Zero-shot Hashing.

Bike Football Cup Bird Computer Car Dicycle Seen Unseen Hashing Model Similar ?

Deep Hashing with Top Similarity Preserving for Image Retrieval

Attribute-Guided Network for Cross-Modal Zero-Shot Hashing

Unsupervised Hashing with Semantic Concept Mining

Deep Hashing Network for Efficient Similarity Retrieval

Self-Distillation Hashing for Efficient Hamming Space Retrieval

Seg-Hashnet: Semantic Segmentation Based Unsupervised Hashing

SSDH: Semi-Supervised Deep Hashing for Large Scale Image Retrieval

Deep Semantic-Preserving and Ranking-Based Hashing for Image Retrieval.

Deep Self-Adaptive Hashing for Image Retrieval

Deep Quantization Network for Efficient Image Retrieval

Hash Learning with Convolutional Neural Networks for Semantic Based Image Retrieval.

Dual-Level Semantic Transfer Deep Hashing for Efficient Social Image Retrieval

Deep Captioning Hashing Network for Complex Scene Image Retrieval

CapsNet-based supervised hashing

Unsupervised Discrete Hashing With Affinity Similarity

Deep Supervised Hashing for Fast Image Retrieval

Learning discrete class-specific prototypes for deep semantic hashing

Deep Hash Learning for Efficient Image Retrieval