2019/07/22 by Kelly Lais Wiggers, Wiggers, Kelly Lais, Alceu de Souza Britto Junior +7
Computer Science · #Advanced Image and Video Retrieval Techniques #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Handwritten Text Recognition Techniques #Image Processing and 3D Reconstruction #Image Retrieval and Classification Techniques #Machine Learning (cs.LG) #Multimedia (cs.MM)
paper · pdf · doi:10.48550/arxiv.1907.09404
openalex publication_date 2019/07/22 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
This paper describes two approaches for content-based image retrieval and\npattern spotting in document images using deep learning. The first approach\nuses a pre-trained CNN model to cope with the lack of training data, which is\nfine-tuned to achieve a compact yet discriminant representation of queries and\nimage candidates. The second approach uses a Siamese Convolution Neural Network\ntrained on a previously prepared subset of image pairs from the ImageNet\ndataset to provide the similarity-based feature maps. In both methods, the\nlearned representation scheme considers feature maps of different sizes which\nare evaluated in terms of retrieval performance. A robust experimental protocol\nusing two public datasets (Tobacoo-800 and DocExplore) has shown that the\nproposed methods compare favorably against state-of-the-art document image\nretrieval and pattern spotting methods.\n