A Framework for Automatic Hair Counting and Measurement

摘要

Alopecia is a research focus in clinical dermatology. Compare with other skin diseases diagnosis, it is a tough task to measure the effects of hair treatment. In the paper, we propose a framework for hair analysis and measurement based on digital image, which includes hair numbers in the region of interested, hair diameter and hair length. Several techniques are considered in our research: an improved classical iterative thresholding method for hair image segmentation motivated by divide and conquer design paradigm, skeleton extraction method for hair counting, and curvature analysis method for cross hair partition. Experiments are performed to demonstrate the effectiveness and robustness of our system.