Iot-based hand vein image enhancement system
Loading...
Date
2019-06
Authors
Teo, Peck Geok
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The low visibility of dorsal hand vein leads to failure in Peripheral Intravenous (PIV) access. This has brought many negative impact to not only patients but to the operator who execute the procedure as well. To aid the PIV access, an effective hand vein enhancement system which can make the vein images more obvious is desired. However, there are limitations in enhancement system developed by previous researcher such as low portability and poor image enhancement ability. This project proposed an enhancement system that combine filtering and enhancement methods from previous undergraduate researchers by using MATLAB software. Raspberry Pi was setup by installing MATLAB hardware support package for Raspberry Pi to enable the interaction between Raspberry Pi hardware and MATLAB software. Different noise levels are modelled to generate noisy image to test the performance of selected filters. Then, Three Value Weighted Filter (TVWF) and Block Matching with 3D (BM3D) filter were applied on the noisy images. The PSNR, SNR, MSE, MSSIM and Q values of the filtered image were calculated. Next, the images produced from best filtering method were enhanced through grayscale enhancement, image segmentation and binary enhancement. Specificity, accuracy and sensitivity of enhanced images were computed. This project adds on the ultilization of IoT application for cloud storage of enhanced hand vein images to ease remote user’s accessibility. GoSync application is installed to Raspberry Pi to connect with Google Drive Cloud for images storage. Lastly, the performance of the proposed system has been compared with the previous work’s results. The results show that TVWF works more effective on hand vein images than BM3D. Both filters worked effective on the smallest noise level, 0.1 for Salt and Pepper Noise and 0.001 for Gaussian Noise. Authorized user can access to the cloud storage anywhere via the Google Drive application.The enhanced images have been stored to the cloud storage. Overall, this work reveals that the proposed image filtering and enhancement system has lower performance as compared to the previous work which employed only image enhancement.