Machine vision in medicine

It is evident from the plethora of articles that AI is being utilized in medicine in a multitude of ways. It is using big data to help determine diagnoses and make decisions more quickly. Just imagine for a moment that you don’t need to communicate with a doctor at all. Describe your feelings. Take a photo. That’s all you have to do. The AI will identify any deviations from the norm and inform the doctor. BOT4.AI has developed an algorithm that identifies deviations and creates an individual training and rehabilitation program to improve health. (Questions: aa@bot4.ai)

This solution will help many people. This solution can be replicated for completely different medical tasks, such as the analysis of rehabilitation or the identification of a primary disease.
The algorithm will analyze three body projections by default: frontal, posterior, and lateral. This allows for the analysis of all possible areas of distortion. The main areas of analysis are the position of the shoulders, elbows, hands, pelvis, knees, feet, and general symmetry of the body.
The system can be implemented in gyms and tell athletes which muscle groups require additional work. It can also be installed in a rehabilitation center and clarify the points of rehabilitation development. BOT4.AI will take care of your health. If you are interested in this implementation, contact us: aa@bot4.ai

Body areas are identified using computer vision algorithms for the following tasks: We detect and segment key points. We used models from the Detectron2 family from Facebook Research, and then we fine-tuned them on our dataset of 10,000 images. The model achieved an accuracy of 0.98 IoU for the segmentation task and MSE=0.001 for the key point determination task.