WCSE

Research on Image Feature Recognition Based on Convolution-Long Short Term Memory Network

Abstract— In order to improve the image recognition rate of Long-Short-Memory-Network, using convolution calculation to reduce dimension and extract feature, which can remove large amounts of redundant information from image samples and accurately extract image features. The recognition rate of image can significantly improve by using classification of serialized image features obtained by compression and dimensionality as the input of LSTM. Firstly, CNN is used to extract the image features accurately. Secondly, the features are serialized into continuous picture bars with strong correlation. LSTM is used to classify the images, thus obtaining a better recognition effect. The experimental results show that the recognition rate of the CNN-LSTM is 30% higher than the basic LSTM, and the recognition rate of the CNN is 5% higher than the basic CNN.

Index Terms— Deep Learning, Long Short Term Memory Network,CNN,cifar-10, Image Recognition.

Chao Yu, Jing Zhou, Liang Gong , Lei Sun , Pengfei Shi, Xinxin Ou
School of Mathematics and Computer Science, Jianghan University, Wuhan Hubei 430056, CHINA.

[Download]

Cite: Chao Yu, Jing Zhou, Liang Gong , Lei Sun , Pengfei Shi, Xinxin Ou, "Research on Image Feature Recognition Based on Convolution-Long Short Term Memory Network," Proceedings of 2019 the 9th International Workshop on Computer Science and Engineering, pp. 176-180, Hong Kong, 15-17 June, 2019.