Self-supervised Product Quantization for Deep Unsupervised Image Retrieval
This addresses the challenge of error-prone and costly labeling for large-scale image retrieval systems, offering an incremental improvement over supervised methods.
The paper tackles the problem of requiring labeled data for deep image retrieval by proposing a self-supervised product quantization method that learns codewords and descriptors without labels, achieving state-of-the-art results on benchmarks.
Supervised deep learning-based hash and vector quantization are enabling fast and large-scale image retrieval systems. By fully exploiting label annotations, they are achieving outstanding retrieval performances compared to the conventional methods. However, it is painstaking to assign labels precisely for a vast amount of training data, and also, the annotation process is error-prone. To tackle these issues, we propose the first deep unsupervised image retrieval method dubbed Self-supervised Product Quantization (SPQ) network, which is label-free and trained in a self-supervised manner. We design a Cross Quantized Contrastive learning strategy that jointly learns codewords and deep visual descriptors by comparing individually transformed images (views). Our method analyzes the image contents to extract descriptive features, allowing us to understand image representations for accurate retrieval. By conducting extensive experiments on benchmarks, we demonstrate that the proposed method yields state-of-the-art results even without supervised pretraining.