Non-contact Venous Imaging Virtual Reality Glasses
Xu Wei, Liang Dingkang, Chen Yihao, Xiao Jian (Mentor), Tong Yi (Mentor)
Nanjing University of Posts and Telecommunications
With the rapid development of computer graphics, the most vivid and convenient way to realize the venous imaging is designed and studied. The non-contact venous display imaging system is mainly based on the infrared image acquisition. Combining visual information with image analysis and dynamic sensing technology, image enhancement provides users with more comfortable venous imaging services.
This work has a wide application prospect. It can help health care workers to find veins, while taking advantage of their own characteristics of non-invasive, nuclear-free medical radiation, reducing the suffering of specific patients and doctor-patient disputes; The complex venous images under human epidermis can be obtained by infrared ray, and the main distribution regions of veins can be identified, which can be used to form venous data template and assist in personal information checking.
Based on PYNQ-Z2, the camera is driven by PL and the motor is controlled. In the linux system, the vein image is segmented by calling the .bit file under pl by overlay, the algorithm is enhanced and the VR glasses are integrated to realize the virtual reality effect of non-contact vein imaging. According to the user's angle of view conversion amplitude control display the corresponding visual field vein image information.
The innovation point of this work is:
1.PYNQ is used to drive the camera and control the motor to process the infrared image efficiently at the same time.
2.Contourlet transform is used to accurately capture the edge contour information of the image to different scales and different directions in the image compression and denoising applications such as image compression and denoising can preserve the texture details information very well.
3.Using the eigenvalue of Hessian matrix as a reference, replacing the traditional method of reference gray value, and combining with the method of comparing the difference of eigenvector of Hessian matrix, we can judge whether a pixel in the image is a venous vessel point or not.
4.combining computer graphics with virtual reality, the non-contact venous imaging is realized, and the imaging effect is more convenient and vivid.