Segmental Deployment of Neural Network in Cloud Robotic System

In this paper, we describe a new method for ep neural networks in the field of computer vision, which can effectively solve the difficulty of applying deep learning in the cloud robotic system. By segmenting the trained network, most of the computing tasks can be cut out and offloaded to the cloud. By effective feature extraction and compression methods, the computing power of robot and cloud can be integrated and coordinated. A method of selecting the split points of the network model and a method of data transmission and compression in the communication between robots and cloud after segmenting are given based on the characteristics of machine vision tasks, and the theoretical analysis is carried out. In the experiment, the effectiveness of all the above methods is verified by comparing the compression capability, response time and network performance of the actual network model. The experimental results show that with the use of segmental methods in cloud robotic system, the task of deep network is processed in real time, while the performance is almost guaranteed.

[1]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[2]  Song Han,et al.  Deep Compression: Compressing Deep Neural Network with Pruning, Trained Quantization and Huffman Coding , 2015, ICLR.

[3]  Xiangyu Zhang,et al.  ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[4]  John R. Gilbert,et al.  Parallel sparse matrix-vector and matrix-transpose-vector multiplication using compressed sparse blocks , 2009, SPAA '09.

[5]  N. Ahmed,et al.  Discrete Cosine Transform , 1996 .

[6]  Jianqin Zhou,et al.  On discrete cosine transform , 2011, ArXiv.

[7]  Mark Sandler,et al.  MobileNetV2: Inverted Residuals and Linear Bottlenecks , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[8]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[9]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.