Title: Evaluating the Impact of computer vision in facial skin disease diagnosis

Abstract:

Facial skin diseases are a major concern among  many people especially in Africa. This comes into play especially as you consider factors such as environment, living standards and a poor disregard for personal skin maintenance which is perhaps vital but the latter group does not share the same views due to a myriad of different thought processes.  With the onset of on device ML algorithms that achieve state of the art performance on different datasets the ideal next step would be to merge facial skin care diagnosis and AI thereby bringing forth a field of AI for skin care. In it’s most primitive form this entails diagnosis as the most reasonable and first achievable step in the pipeline of solving this maze. The major reason to this is mainly because of the onset of powerful handsets and I would think to be structured data collection and skilled people who annotate this same data. Also more important is that very people can access dermatologists and if they do it mostly as a reactive response in that something really bad has to happen to compel a visit to one.

 The major people I spoke to wee mostly dermatologists and also people whom found it resourceful to share their different opinions via structured one on one interview as I though it to be a very sensitive topic. The most important bit of it was especially that most skin diseases were diagnosed mainly by the visual periphery and so dermatologist's need to have a keen eye to be able to spot this. This is where computer vision comes in and especially transformers which yield a very large potential in trying to diagnose this different ailments. Some of the skin diseases diagnosable via this process include, acne, atopic dermatitis, black spots, warts, dry skin, oily skin and fungal infection on the face. It offers a wide range of selection fo different possibilities so the best idea is to maybe pick the ones that have a most reoccurrence rate and are much easily to  spot for initial use cases whereby the model is not most powerful yet. This would majorly apply to women whom in their own rights look after their skin more in comparison to men and so even data collection would be favored by this metric. Improvement is majorly possible because a data flywheel approach is implemented.

Biography:

He is passional in innovation particularly in the technology spectrum whereby I am focused on building intelligent systems using machine learning and can use data to its fullest. He is skilled in Python and JavaScript and he have mastered the various frameworks. He have mastered Django, Flask and also Containerization using Docker alongside React which is a JavaScript framework. He is more enlightened in solving problems using code using data whereby he can leverage his most prominent skills in Machine Learning.