Abstract
Image retrieval techniques are becoming famous due to the vast availability of multimedia data. The present image retrieval system performs excellently on labeled data. However, often, data labeling becomes costly and sometimes impossible. Therefore, self-supervised and unsupervised learning strategies are currently becoming illustrious. Most of the self/unsupervised strategies are sensitive to the number of classes and can not mix labeled data on availability. In this paper, we introduce AutoRet, a deep convolutional neural network (DCNN) based self-supervised image retrieval system. The system is trained on pairwise constraints. Therefore, it can work in self-supervision and can also be trained on a partially labeled dataset. The overall strategy includes a DCNN that extracts embeddings from multiple patches of images. Further, the embeddings are fused for quality information used for the image retrieval process. The method is benchmarked with three different datasets. From the overall benchmark, it is evident that the proposed method works better in a self-supervised manner. In addition, the evaluation exhibits the proposed method’s performance to be highly convincing while a small portion of labeled data are mixed on availability.
Highlights
Due to the explosion of smartphones and social media, the quantity of image-based information is rapidly increasing
We compare our model with the following un/self-supervised models: sparse graph based self-supervised hashing (SGSH) [18], self-supervised product quantization (SPQ) [14], deep variational binaries (DVB) [55], distillhash [17], binary generative adversarial networks (BGAN) [31], BinGAN [56], unsupervised deep hashing with pseudo labels (UDHP) [20], similarity adapdive deep hashing (SADH) [16]
The paper proposes an image retrieval system, AutoRet, which can establish image relationships based on image content
Summary
Due to the explosion of smartphones and social media, the quantity of image-based information is rapidly increasing. CBIR is currently dominating due to the heavy requirement of image-based information retrieval systems. This paper introduces a general-purpose image retrieval system based on self-supervised learning. Self-supervised learning strategies are trained with pseudo-labeled data. The current semi-supervised systems are limited to hash-based retrieval methods. Hash-based retrieval methods use DCNN as a hash function to generate a binary representation for a given image. Most algorithms focus on self/semi/un-supervised learning strategies, neglecting the process of a partially labeled dataset. This paper introduces a self-supervised general-purpose image retrieval system with some advantages. We introduce a CBIR system named AutoRet, which can be trained in self-supervised and can be integrated with labeled data as well. We evaluate our model with different image retrieval techniques involving self/unsupervised strategies and validate that AutoRet performs better in all scenarios.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
