Adaptive Feature Extraction Using Sparse Autoencoder Based On Bag-of-Word Model
Krit Chaiso, Jaktip Yodsri, Kanjanapan Sukvichai · ITC-CSCC :International Technical Conference on Circuits Systems, Computers and Communications · 2015
In computer vision, recognition, the presence of the surrounding objects, is one of the most important and challenging task. Purposing the deep-learning approach is to accomplish these important tasks, in this research, the autoencoder technique and Bag-of- Word model (BoW) are thoroughly used in order to obtain the practically meaningful extracted features as well as the reliable classification engine. Conducting the experiment by applying this approach to the recognition tasks, it shows the good result in the prediction accuracy. The results shows that appropriate dimension of network structure produced acceptable features and better performance.