DRFNet: A lightweight and high accuracy network for resource-limited implementation
Hengliang Xiang, Liang Chen, Wei Hong Xu · 2017
With the increased demand of recognition accuracy in image classification, the convolutional neural networks (CNNs) grow rapidly in scale, which makes CNNs more difficult to be deployed on resource-limited platforms. In this paper, we present a lightweight Deep Residual FireNet (DRFNet). The new DRFNet has 26 weight layers and uses the fire module as its basic building module. In its architecture, we use three max-pooling operations, and additionally two average-pooling layers to preserve a higher resolution and larger receptive field. In addition, four bypass connections are used to further improve the accuracy rate. With this new design, the proposed DRFNet has successfully reduced the number of weight parameters to be around 1.85 million, 3% of AlexNet and 1.3% of VGG16. Meanwhile, the new DRFNet has achieved 91.44% top-5 accuracy rate on the ImageNet validation dataset, which is higher than AlexNet, SqueezeNet, VGG16 and GoogLeNet.