{"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/an-architecture-combining-convolutional","title":"An Architecture Combining Convolutional Neural Network (CNN) and Support Vector Machine (SVM) for Image Classification","arxiv_id":"1712.03541","date":"2017-12-10","proceeding":null,"authors":["Abien Fred Agarap"],"abstract":"Convolutional neural networks (CNNs) are similar to \"ordinary\" neural\nnetworks in the sense that they are made up of hidden layers consisting of\nneurons with \"learnable\" parameters. These neurons receive inputs, performs a\ndot product, and then follows it with a non-linearity. The whole network\nexpresses the mapping between raw image pixels and their class scores.\nConventionally, the Softmax function is the classifier used at the last layer\nof this network. However, there have been studies (Alalshekmubarak and Smith,\n2013; Agarap, 2017; Tang, 2013) conducted to challenge this norm. The cited\nstudies introduce the usage of linear support vector machine (SVM) in an\nartificial neural network architecture. This project is yet another take on the\nsubject, and is inspired by (Tang, 2013). Empirical data has shown that the\nCNN-SVM model was able to achieve a test accuracy of ~99.04% using the MNIST\ndataset (LeCun, Cortes, and Burges, 2010). On the other hand, the CNN-Softmax\nwas able to achieve a test accuracy of ~99.23% using the same dataset. Both\nmodels were also tested on the recently-published Fashion-MNIST dataset (Xiao,\nRasul, and Vollgraf, 2017), which is suppose to be a more difficult image\nclassification dataset than MNIST (Zalandoresearch, 2017). This proved to be\nthe case as CNN-SVM reached a test accuracy of ~90.72%, while the CNN-Softmax\nreached a test accuracy of ~91.86%. The said results may be improved if data\npreprocessing techniques were employed on the datasets, and if the base CNN\nmodel was a relatively more sophisticated than the one used in this study.","url_abs":"http://arxiv.org/abs/1712.03541v2","url_pdf":"http://arxiv.org/pdf/1712.03541v2.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":"an-architecture-combining-convolutional","repo_url":"https://github.com/AFAgarap/cnn-svm","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"an-architecture-combining-convolutional","repo_url":"https://github.com/da-moon/classifiers-monorepo","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1712.03541","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}