A low-cost implementation method on deep neural network using stochastic computing

2021 ◽  
Author(s):  
Ya Dong ◽  
Xingzhong Xiong ◽  
Tianyu Li ◽  
Lin Zhang ◽  
Jienan Chen
Keyword(s):  
Neural Network ◽  
Deep Neural Network ◽  
Low Cost ◽  
Stochastic Computing ◽  
Implementation Method

scholarly journals RGB-D Object Recognition Using Multi-Modal Deep Neural Network and DS Evidence Theory

Sensors ◽  
2019 ◽  
Vol 19 (3) ◽  
pp. 529 ◽  
Author(s):  
Hui Zeng ◽  
Bin Yang ◽  
Xiuqing Wang ◽  
Jiwei Liu ◽  
Dongmei Fu
Keyword(s):  
Neural Network ◽  
Object Recognition ◽  
Deep Neural Network ◽  
Low Cost ◽  
Evidence Theory ◽  
Feature Learning ◽  
Decision Fusion ◽  
Support Vector ◽  
Depth Sensors ◽  
Depth Images

With the development of low-cost RGB-D (Red Green Blue-Depth) sensors, RGB-D object recognition has attracted more and more researchers’ attention in recent years. The deep learning technique has become popular in the field of image analysis and has achieved competitive results. To make full use of the effective identification information in the RGB and depth images, we propose a multi-modal deep neural network and a DS (Dempster Shafer) evidence theory based RGB-D object recognition method. First, the RGB and depth images are preprocessed and two convolutional neural networks are trained, respectively. Next, we perform multi-modal feature learning using the proposed quadruplet samples based objective function to fine-tune the network parameters. Then, two probability classification results are obtained using two sigmoid SVMs (Support Vector Machines) with the learned RGB and depth features. Finally, the DS evidence theory based decision fusion method is used for integrating the two classification results. Compared with other RGB-D object recognition methods, our proposed method adopts two fusion strategies: Multi-modal feature learning and DS decision fusion. Both the discriminative information of each modality and the correlation information between the two modalities are exploited. Extensive experimental results have validated the effectiveness of the proposed method.

scholarly journals Classification of 41 Hand and Wrist Movements via Surface Electromyogram Using Deep Neural Network

2021 ◽  
Vol 9 ◽  
Author(s):  
Panyawut Sri-iesaranusorn ◽  
Attawit Chaiyaroj ◽  
Chatchai Buekban ◽  
Songphon Dumnin ◽  
Ronachai Pongthornseri ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Neural Network ◽  
Low Cost ◽  
Sampling Rate ◽  
Assessment Procedure ◽  
Semg Signal ◽  
Neural Network Approach ◽  
Assessment Protocol ◽  
Wide Range

Surface electromyography (sEMG) is a non-invasive and straightforward way to allow the user to actively control the prosthesis. However, results reported by previous studies on using sEMG for hand and wrist movement classification vary by a large margin, due to several factors including but not limited to the number of classes and the acquisition protocol. The objective of this paper is to investigate the deep neural network approach on the classification of 41 hand and wrist movements based on the sEMG signal. The proposed models were trained and evaluated using the publicly available database from the Ninapro project, one of the largest public sEMG databases for advanced hand myoelectric prosthetics. Two datasets, DB5 with a low-cost 16 channels and 200 Hz sampling rate setup and DB7 with 12 channels and 2 kHz sampling rate setup, were used for this study. Our approach achieved an overall accuracy of 93.87 ± 1.49 and 91.69 ± 4.68% with a balanced accuracy of 84.00 ± 3.40 and 84.66 ± 4.78% for DB5 and DB7, respectively. We also observed a performance gain when considering only a subset of the movements, namely the six main hand movements based on six prehensile patterns from the Southampton Hand Assessment Procedure (SHAP), a clinically validated hand functional assessment protocol. Classification on only the SHAP movements in DB5 attained an overall accuracy of 98.82 ± 0.58% with a balanced accuracy of 94.48 ± 2.55%. With the same set of movements, our model also achieved an overall accuracy of 99.00% with a balanced accuracy of 91.27% on data from one of the amputee participants in DB7. These results suggest that with more data on the amputee subjects, our proposal could be a promising approach for controlling versatile prosthetic hands with a wide range of predefined hand and wrist movements.

scholarly journals Some New Perspectives on the Multi-user Detection in Uplink Grant-Free NOMA using Deep Neural Network

2021 ◽  
Author(s):  
Shah Mahdi Hasan ◽  
Kaushik Mahata ◽  
Md Mashud Hyder
Keyword(s):  
Neural Network ◽  
Deep Neural Network ◽  
Low Cost ◽  
Base Station ◽  
Data Detection ◽  
Free System ◽  
Active User ◽  
Important Concern ◽  
Wide Range ◽  
Alternative Routes

To support the explosive growth of the Internet of Things (IoT), Uplink (UL) grant-free Non-Orthogonal Multiple Access (NOMA) emerges as a promising technology. It has the potential of offering scalable and low-cost solutions for the resource-constrained Massive Machine Type Communication (mMTC) systems. In principle, the grant-free NOMA enables small signaling overhead and low access latency time by circumventing complicated grant-access based procedures which is commonly found in the legacy wireless networks. In a UL grant-free system, a complete Multi-User Detection (MUD) algorithm not only performs the Active User Detection (AUD) but also the Channel Estimation (CE) and the Data Detection (DD). By exploiting the naturally occurring sparse user activity in the mMTC systems, the MUD problem can be solved using a wide range of Compressive Sensing based algorithms (CS-MUD). However, some alternative routes have been explored in the literature as well. The utility of these algorithms, in general, revolve around some assumptions about the channel or the availability of perfect channel information at the Base Station (BS). How these assumptions are met in a practical circumstance is, however, an important concern. In this work we devise an end-to-end MUD using Deep Neural Network (DNN) where we relax these assumptions. We approximate an ensemble of trained DNN based MUD using Knowledge Distillation (KD) to enable fast AUD at the Base Station (BS). Furthermore, using the inter-resource correlation, we estimate the channels of the active users which is an ill-posed problem otherwise. We carry out elaborate numerical investigation to validate the efficacy of the proposed approach for the UL grant-free NOMA systems.

Sign in / Sign up

Export Citation Format

Share Document