Title :
A discriminative cascade CNN model for offline handwritten digit recognition
Author :
Shulan Pan ; Yanwei Wang ; Changsong Liu ; Xiaoqing Ding
Author_Institution :
Dept. of Electron. Eng., Tsinghua Univ., Beijing, China
Abstract :
This paper presents a high-performance two-stage cascade CNN model. The main idea behind the cascade CNN model is complementary classification objectives between Stage I and Stage II. Discriminative learning is introduced to train Stage II by feeding back poorly recognized training samples. Experiments have been conducted on the competitive MNIST handwritten digit database. The cascade model achieved the best state-of-the-art performance with an error rate of 0.18%.
Keywords :
convolution; error analysis; handwritten character recognition; learning (artificial intelligence); neural nets; MNIST handwritten digit database; complementary classification objective; convolutional neural network; discriminative cascade CNN model; discriminative learning; error rate; offline handwritten digit recognition; training sample; Distortion; Error analysis; Handwriting recognition; Neural networks; Support vector machines; Training; Writing;
Conference_Titel :
Machine Vision Applications (MVA), 2015 14th IAPR International Conference on
Conference_Location :
Tokyo
DOI :
10.1109/MVA.2015.7153240
