Skin Disease Detection - Abto Software

Today, it is said that there are more than 125 million people who suffer from psoriasis, the number of skin cancer cases is growing each year. Key factor in skin conditions treatment is their early detection, since many skin diseases, if not treated, can cause further health complications in the body including spreading the infection from one individual to the other, or even be fatal like melanoma skin cancer.

Developing an efficient medical imaging solution for skin disease detection poses quite a few challenges derived mostly from the diversity of visual skin disease characteristics. Skin undertone, skin color, even skin texture and coarseness, all that play very important role in skin disease detection, since they all make the same disease show itself differently.

Our skin disease detection solution uses digital image processing techniques for the classification of infected skin.

illustration of skin disease

How it works?

An automatic medical images analysis system can be divided into three stages:

proper enhancement
feature extraction and selection
classification

The proper enhancement is the most important, since it affects the precision of the next steps. Image Enhancement is the process of adjusting digital image so that the result is more suitable for further medical image analysis. It is achieved by removing noise, sharpening or brightening the medical images, thus making it easier to identify key features.

LEARN MORE

The unique features of the enhanced medical images are extracted using HSV histogram and SURF. We then classify the images, based on the features, as infected skin and normal skin. This methodology provides good accuracy. For the next version we plan to vary the image processing techniques and classifiers, so as to get even higher results.

The data we have gathered may be used in medical science for the development of specific skin disease recognition systems used as a telemedicine tool to detect all the infected areas on the skin, including those that are invisible to human eyes.

Medical images collection
Enhancement of medical images
Feature Extraction using HSV Histogram and SURF
Define the group and class of medical images
kNN-classify ( group, class )

Proposed Solution framework for Skin Disease Detection

Features of Skin Disease Detection Solution

Machine learning powered algorithm
Our skin disease detection prototype uses machine learning and computer vision techniques to achieve the highest accuracy, and this kind of system improves with the expansion of our database.
You can monitor your health 24/7
Even though it’s still a prototype, we think, it’s a step in the right direction of the next-generation telemedicine to help you keep a closer, more informed eye on the health of your skin.
AI for custom e-health software development
Our detection algorithms can be applied to develop a simple, inexpensive software system to identify specific skin disease, even detect and recognize stages of skin cancer.