A Survey of Image Captioning Algorithms Based on Deep Learning
Fan Yang, Xueping Su, Jie Ren, Xiaomin Ma, Yongyong Han · 2022
As a cross task of natural language processing and computer vision, image captioning is a key technology to explore the transformation of artificial intelligence visual perception to high-level semantic understanding. It is widely used in image retrieval, blind navigation, early childhood education, human-computer interaction, safe assisted driving and other fields. Firstly, the research progress of image captioning is briefly described, and then the image captioning algorithms based on deep learning methods are classified and summarized according to their development history, and the main ideas and core issues of various algorithms are elaborated. Finally, the implementation results of various algorithms are compared, the problems existing in the existing algorithms are summarized and analyzed, and the future development trend of image captioning is prospected.