Softmax Cross Entropy Loss with Unbiased Decision Boundary for Image Classification
Authors:
- 1 November 2018
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2028-2032
Abstract
Considering that in neural network based on softmax cross entropy loss, the output probability is mainly based on linear computation of parameter vectors of each class in the last layer and hidden features in the layer of sample points. Therefore, the final output of neural network is effected by of the L2-norm of parameter vector of each class. Taking binary-class as an example, if the parameter vector of a class has a large L2-norm., decision boundary is close to another class with smaller L2-norm., so that sample points will be easily assigned to the class with large L2-norm. Based on it, this paper proposes a new softmax cross entropy loss, which adjusts the position of decision boundary so that it is not biased to any class. Experimental results on the LabelMe dataset and the UIUC-Sports dataset show that the proposed loss is superior to softmax cross entropy loss.Keywords
This publication has 8 references indexed in Scilit:
- Image-Text Dual Model for Small-Sample Image Classification2017
- Focal Loss for Dense Object Detection2017
- Improvement of learning for CNN with ReLU activation by sparse regularization2017
- Decorrelation of Neutral Vector Variables: Theory and ApplicationsIEEE Transactions on Neural Networks and Learning Systems, 2016
- PCANet: A Simple Deep Learning Baseline for Image Classification?IEEE Transactions on Image Processing, 2015
- FaceNet: A unified embedding for face recognition and clustering2015
- Spatial Pyramid Pooling in Deep Convolutional Networks for Visual RecognitionIEEE Transactions on Pattern Analysis and Machine Intelligence, 2015
- Face recognition: a convolutional neural-network approachIEEE Transactions on Neural Networks, 1997