Exploring Deep Learning Approach on Semantic Gap: A Comprehensive Review

Abstract

With the prevalence of the Internet and smartphones, users upload a large number of images to the web. However, it is challenging for users to find what they really need from the vast sea of images. It is also difficult for Internet companies to effectively integrate their massive image data resources. In traditional content-based image retrieval, images are indexed by their low-level visual features, which leads to a key problem: the semantic gap between low-level features and high-level semantic concepts. To address this issue, semantic-based image retrieval has been proposed as a solution to bridge the semantic gap, making it a key technical challenge in the field of image retrieval. To tackle these challenges, this proposal presents a novel multi-annotation method for images and develops an image retrieval system based on deep learning and image semantic content. Preparatory work, including a literature review and methodology development, will be conducted to implement the semantic-based image retrieval system and efficiently utilize the vast number of available images.