Bag-of-visual-words for signature-based multi-script document retrieval

摘要: An end-to-end architecture for multi-script document retrieval using handwritten signatures is proposed in this paper. The user supplies a query signature sample, and the system exclusively returns set of documents that contain signature. In first stage, component-wise classification technique separates potential components from all other components. A bag-of-visual-words powered by SIFT descriptors patch-based framework to compute features support vector machine (SVM)-based classifier was used separate documents. second foreground (i.e., strokes) background spatial information loops, reservoirs etc.) were combined characterize object match with Finally, three distance measures present target retrieval. ‘Tobacco’ (The Legacy Tobacco Document Library (LTDL). University California, San Francisco, 2007. http://legacy.library.ucsf.edu/) database an Indian script containing 560 Devanagari (Hindi) Bangla scripts performance evaluation. also tested on noisy documents, promising results obtained. comparative study shows method outperforms state-of-the-art approaches.