A novel hybrid CNN–SVM classifier for recognizing handwritten digits

• Published in: Pattern recognition Vol. 45; no. 4; pp. 1318 - 1325
• Main Authors: Niu, Xiao-Xiao, Suen, Ching Y.
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 01.04.2012; Elsevier
• Subjects: Applied sciences; Artificial intelligence; Classifiers; Computer science; control theory; systems; Connectionism. Neural networks; Convolutional Neural Network; Digits; Exact sciences and technology; Handwritten digit recognition; Hybrid model; Information, signal and communications theory; Mathematical models; Pattern recognition; Recognition; Rejection; Signal and communications theory; Signal processing; Signal representation. Spectral analysis; Signal, noise; Support Vector Machine; Support vector machines; Telecommunications and information theory; Convolutional Neural Network; Hybrid model; Support Vector Machine; Handwritten digit recognition; Performance evaluation; Automatic classification; Cellular neural nets; Support vector machine; Pattern recognition; Neural network; Signal classification; Database; Handwritten character recognition; Manuscript character
• ISSN: 0031-3203; 1873-5142
• DOI: 10.1016/j.patcog.2011.09.021

This paper presents a hybrid model of integrating the synergy of two superior classifiers: Convolutional Neural Network (CNN) and Support Vector Machine (SVM), which have proven results in recognizing different types of patterns. In this model, CNN works as a trainable feature extractor and SVM performs as a recognizer. This hybrid model automatically extracts features from the raw images and generates the predictions. Experiments have been conducted on the well-known MNIST digit database. Comparisons with other studies on the same database indicate that this fusion has achieved better results: a recognition rate of 99.81% without rejection, and a recognition rate of 94.40% with 5.60% rejection. These performances have been analyzed with reference to those by human subjects. ► We explored a new hybrid of Convolutional Neural Network and Support Vector Machine. ► Experiments were conducted on the MNIST database. ► The hybrid model has achieved better recognition and reliability performances. ► The best recognition rate was 99.81% without rejection. ► A reliability rate of 100% with 5.60% rejection was obtained.