{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/open-source-dataset-and-deep-learning-models","title":"Open Source Dataset and Deep Learning Models for Online Digit Gesture Recognition on Touchscreens","arxiv_id":"1709.06871","date":"2017-09-20","proceeding":null,"authors":["Philip J. Corr","Guenole C. Silvestre","Chris J. Bleakley"],"abstract":"This paper presents an evaluation of deep neural networks for recognition of\ndigits entered by users on a smartphone touchscreen. A new large dataset of\nArabic numerals was collected for training and evaluation of the network. The\ndataset consists of spatial and temporal touch data recorded for 80 digits\nentered by 260 users. Two neural network models were investigated. The first\nmodel was a 2D convolutional neural (ConvNet) network applied to bitmaps of the\nglpyhs created by interpolation of the sensed screen touches and its topology\nis similar to that of previously published models for offline handwriting\nrecognition from scanned images. The second model used a 1D ConvNet\narchitecture but was applied to the sequence of polar vectors connecting the\ntouch points. The models were found to provide accuracies of 98.50% and 95.86%,\nrespectively. The second model was much simpler, providing a reduction in the\nnumber of parameters from 1,663,370 to 287,690. The dataset has been made\navailable to the community as an open source resource.","url_abs":"http://arxiv.org/abs/1709.06871v1","url_pdf":"http://arxiv.org/pdf/1709.06871v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"open-source-dataset-and-deep-learning-models","repo_url":"https://github.com/PhilipCorr/numeral-gesture-dataset","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"gesture-recognition","task_name":"Gesture Recognition"},{"task_slug":"handwriting-recognition","task_name":"Handwriting Recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}