Automatic Detection of Natural Skin Markers
Automatic detection of natural skin markers for the early detection of skin cancer in full body imaging.
Category: Computer Assisted Diagnosis
Technologies: Image Processing.
This module is a essential component of a suite for the detection of skin cancer. The developed algorithms contributes to the automation of a procedure for medical diagnosis in whole body photo imaging, which are used in the early detection of potentially deadly skin cancers.
The purpose of this module is the automatic detection of natural skin markers. Its design is based on the metaphor of a spider lands on the image, as shown in the following figure.
The spider at the right of the figure is specialized in the recognition of natural skin markers. This is a difference with the spiders from the natural world (left image), which are specialized in the recognition of insects. When we place our spider on an image of 2112×2816 pixels, such as those produced by our client, our spider visits the 6 million pixels and determines whether each pixel belongs to a natural skin marker. In the case of an 8-legged spider as shown in figure, she performs a total of 8 + 9 evaluations per pixel, which in total represent 100 million evaluations per image, while a 32-legged spider makes 32 + 9 evaluations per pixel, which represents 240 million evaluations per image. This is done by our spider very quickly at the touch of a button.
The following figure shows the results of the spider process, for two images which were taken over time.
Elastic Registration of two Images Spaced in Time using Trac...
Elastic Registration of two images using Natural Skin Marker...