Microfluidic paper-based analytical devices can beused in preliminary screening and diagnosis of specific diseases.It has the advantages of low costs and quick diagnosis. Afterchemical reaction, the paper chip will display the disease'srelative status through color identification or electrochemistrytesting. When compared to electrochemistry testing, the noncontactcolor identification method has the advantages of low
costs, reusability and not requiring cleaning. However sinceresearch in the past have encountered many roblems with thecolor identification method, this has led to the negation of theadvantages of microfluidic paper-based analytical devices inquick testing. For this reason, this study developed a portabletesting apparatus with image analysis as a foundation and uses aself-designed imaging processing calculation method toimplement color identification. The hardware equipmentdesigned for image processing in this study uses a microembedded style system and imaging equipment as the hardwareframework, along with a Web application to implement a multiplatformoperating interface. Color quantization was used as thefoundation for the image processing method, to calculate thepaper chip's color. Furthermore, this study also carried out avariety of color quantization methods to compare the colorquantization results of the icrofluidic paper-based analyticaldevices. Methods include uniform color quantization, median-cutalgorithm, k-means clustering algorithm and self-organizingmaps neural networks. During the experiment phase of this study,we found that median-cut was superior to other methods inapplication. This result is different from many other researchworks conducted on color quantization. Lastly, the experiment
results showed the usability of this testing setup
