Detect the colors of tongue
Participated in building an Artificial Intelligence aided diagnosis mobile web application
- design functions implement image segmentation by using clustering methods (Fuzzy c mean & K means).
- Recognized color and texture features of patients’ tongues and generate the diagnostic reports with machine learning (SVM) and convolutional neural network (Inception V4) on TensorFlow.
This is project that I use computer vision techniques to identify the color and texture features of patients’ tongues, providing a quick diagnostic aid. I have written a function which is able to recognize the “feature” within 20 milliseconds based on the predicted patterns of over 50,000 pixels per picture by the SVM classifier, and gained much hands-on experience in tools/frameworks (e.g. TensorFlow) for image processing, as well as a variety of classifications and clustering algorithms (e.g. R-CNN, XGBoost, FCM, etc.).