Design and implementation of a single-frame super-resolution processing alogrithm for lensless microfluidic imaging system
Date of Issue2015
School of Electrical and Electronic Engineering
The contact-image based microfluidic cytometer with machine-learning for single-frame super-resolution processing is used to count and recognize multiple types of cells in clinical diagnostics and biological research As an extension research, this paper proposed a new algorithm that combines external example-based sparse representation and convolutional neural network. Compared with extreme learning machine algorithm in terms of PSNR, SSIM and running time, the proposed algorithm can perform better in recovering high-resolution images at the expense of more running time. In addition, with the aim to running this system in FPGA, the ELM-SR should be programmed in Verilog. The first step is to transferring the MATLAB codes to C++ code, which has been done in this paper. The result is satisfactory and the difference is analysed in terms of MATLAB and C++ property. As an extension research, this paper not only explored more deeply based on previous job, also made a good preparation for future work.
DRNTU::Engineering::Electrical and electronic engineering
Final Year Project (FYP)
Nanyang Technological University