ThreshNet: An Efficient DenseNet Using Threshold Mechanism to Reduce Connections
學年 111
學期 1
出版(發表)日期 2022-08-11
作品名稱 ThreshNet: An Efficient DenseNet Using Threshold Mechanism to Reduce Connections
作品名稱(其他語言)
著者 Rui-Yang Ju, Ting-Yu Lin, Jia-Hao Jian, Jen-Shiun Chiang, and Wei-Bin Yang
單位
出版者
著錄名稱、卷期、頁數 IEEE Access 10, p.82834-82843
摘要 With the continuous development of neural networks for computer vision tasks, more and more network architectures have achieved outstanding success. As one of the most advanced neural network architectures, DenseNet shortcuts all feature maps to solve the model depth problem. Although this network architecture has excellent accuracy with low parameters, it requires an excessive inference time. To solve this problem, HarDNet reduces the connections between the feature maps, making the remaining connections resemble harmonic waves. However, this compression method may result in a decrease in the model accuracy and an increase in the parameters and model size. This network architecture may reduce the memory access time, but its overall performance can still be improved. Therefore, we propose a new network architecture, ThreshNet, using a threshold mechanism to further optimize the connection method. Different numbers of connections for different convolution layers are discarded to accelerate the inference of the network. The proposed network has been evaluated with image classi_cation using CIFAR 10 and SVHN datasets under platforms of NVIDIA RTX 3050 and Raspberry Pi 4. The experimental results show that, compared with HarDNet68, GhostNet, MobileNetV2, Shuf_eNet, and Ef_cientNet, the inference time of the proposed ThreshNet79 is 5%, 9%, 10%, 18%, and 20% faster, respectively. The number of parameters of ThreshNet95 is 55% less than that of HarDNet85. The new model compression and model acceleration methods can speed up the inference time, enabling network models to operate on mobile devices.
關鍵字 Mobile platform, Raspberry Pi, convolutional neural network, image classication, model compression, model acceleration, threshold mechanism
語言 en_US
ISSN
期刊性質 國外
收錄於 SCI
產學合作
通訊作者 Jen-Shiun Chiang
審稿制度
國別 USA
公開徵稿
出版型式 ,電子版,紙本
SDGS 優質教育,產業創新與基礎設施