Translator Disclaimer
21 March 2014 Context based algorithmic framework for identifying and classifying embedded images of follicle units
Author Affiliations +
Medical image processing has been very emerging research areas in recent days. These types of images are naturally so noisy. To count the target objects is never easy. But the proper treatment depends on the accuracy of the successful locating and counting of the desired objects in an image. Some research work can do this type of segmentation of images, but they include so many constraints on the input images that these solutions cannot be applied in a generalized way to most of the images. Even a slight variation in nature of an input image can lead to a major incorrectness of the result. So we developed a generalized way to count a very noisy part of human body, the hair follicle on the scalp. The objective of this research is to count the number of hair follicle groups and the number of follicles into each group in a microscopic image of human scalp. The follicles are nonstandard in shape i.e. they do not maintain any standard shape like rectangle, oval, circle etc. Moreover the follicles are overlapping with one another in many cases. So it is hard to separate them. Here we will present a technique to count the number of follicle group as well as number of follicles in each group. We also applied well-known techniques to cluster the objects detected and a method to generate a neighboring connected graph in order to calculate the inter follicular distances.
© (2014) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Md. Mahbubur Rahman, S. S. Iyengar, Wei Zeng, Frank Hernandez, Bernard P. Nusbaum, and Paul Rose "Context based algorithmic framework for identifying and classifying embedded images of follicle units", Proc. SPIE 9034, Medical Imaging 2014: Image Processing, 903439 (21 March 2014);

Back to Top