Improving CNN Performance Accuracies With Min-Max Objective
Improving CNN Performance Accuracies With Min-Max Objective
复制标题
使用最小-最大目标提高 CNN 性能准确性
DOI：
10.1109/tnnls.2017.2705682
复制
发表时间：
2017
影响因子：
10.4
通讯作者：
Zheng Nanning
中科院分区：
文献类型：
--
作者：
Shi Weiwei;Gong Yihong;Tao Xiaoyu;Wang Jinjun;Zheng Nanning
我们提出了一种新颖的方法来提高卷积神经网络（CNN）的性能准确性，而无需增加网络复杂性。我们通过在训练过程中将所提出的最小-最大目标应用于 CNN 模型输出层下方的一层来实现该目标。最小-最大目标明确确保 CNN 模型学习的特征图具有每个对象流形的最小流形内距离以及不同对象流形之间的最大流形间距离。最小-最大目标是通用的，能够应用于不同的 CNN，计算成本的增加微乎其微。此外，还提出了与最小-最大目标结合的增量小批量训练程序，以能够处理大规模训练数据。对图像分类和人脸验证任务的多个基准数据集进行的综合实验评估表明，与不使用该目标训练的相同模型相比，在训练过程中采用所提出的最小-最大目标可以显着提高 CNN 模型的性能准确性。
We propose a novel method for improving performance accuracies of convolutional neural network (CNN) without the need to increase the network complexity. We accomplish the goal by applying the proposed Min-Max objective to a layer below the output layer of a CNN model in the course of training. The Min-Max objective explicitly ensures that the feature maps learned by a CNN model have the minimum within-manifold distance for each object manifold and the maximum between-manifold distances among different object manifolds. The Min-Max objective is general and able to be applied to different CNNs with insignificant increases in computation cost. Moreover, an incremental minibatch training procedure is also proposed in conjunction with the Min-Max objective to enable the handling of large-scale training data. Comprehensive experimental evaluations on several benchmark data sets with both the image classification and face verification tasks reveal that employing the proposed Min-Max objective in the training process can remarkably improve performance accuracies of a CNN model in comparison with the same model trained without using this objective.