2010/03/09 by Nibaran Das, Das, Nibaran, Ayatullah Faruk Mollah +5 · 1 citation
Computer Science · Engineering · #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 #Vehicle License Plate Recognition
paper · pdf · doi:10.48550/arxiv.1003.1894
openalex publication_date 2010/03/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The work presents a comparative assessment of seven different feature sets\nfor recognition of handwritten Arabic numerals using a Multi Layer Perceptron\n(MLP) based classifier. The seven feature sets employed here consist of shadow\nfeatures, octant centroids, longest runs, angular distances, effective spans,\ndynamic centers of gravity, and some of their combinations. On experimentation\nwith a database of 3000 samples, the maximum recognition rate of 95.80% is\nobserved with both of two separate combinations of features. One of these\ncombinations consists of shadow and centriod features, i. e. 88 features in\nall, and the other shadow, centroid and longest run features, i. e. 124\nfeatures in all. Out of these two, the former combination having a smaller\nnumber of features is finally considered effective for applications related to\nOptical Character Recognition (OCR) of handwritten Arabic numerals. The work\ncan also be extended to include OCR of handwritten characters of Arabic\nalphabet.\n