Open AccessImbalanced Thangka Image Classification research Based on the ResNet Network
Author(s) -
Fuliang Zeng,
Wenjin Hu,
Guoyuan He,
Chengyu Yue
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1748/4/042054
Subject(s) - softmax function , cross entropy , computer science , artificial intelligence , training set , complement (music) , pattern recognition (psychology) , information loss , algorithm , artificial neural network , convergence (economics) , entropy (arrow of time) , contextual image classification , data set , backpropagation , image (mathematics) , biochemistry , chemistry , physics , quantum mechanics , complementation , economics , gene , phenotype , economic growth
Aiming at the problem of performance degradation caused by ignoring the softmax score of the wrong class in the training process of the unbalanced data set of Thangka images and the problem of the loss of negative feature information in the propagation process of the ReLU activation function, a new loss calculation method is proposed. Firstly, the parameters of the pre-training model on the COCO data set are used as the initial parameters. Secondly, CE and CCE are used to calculate the loss during the calculation of loss in the back propagation. Finally, AReLU activation function is used and a weight assigned to CE and CCE is added as final loss to update the parameters. The experimental results show that this algorithm improves the convergence speed and accuracy of the model with respect to imbalanced data. Compared with other loss functions, ours method performance is state-of-the-art, such as complement cross entropy.