Visual patterns recognition, computer vision and tracking of objects by video on the example of detecting the direction of the newborn's view and analysis of their visual activity
We are actively developing a number of solutions for non-invasive diagnosis of various pathological conditions of newborns using video analysis.
For an objective evaluation of the methods of visual stimulation of prematurity, we developed a system for detecting the child's view based on the position of the pupils. Using the methods of computer vision and machine learning, the program determines with good accuracy the direction of view on the video, shot by the usual modern smartphone
Analyzing the reaction time of the child, the degree of involvement, the interest of the child and how the child looks at the pictures that is shown to him we try to follow the development of neurosensory abilities of premature babies and try to give an objective assessment of the quality of the applied methods of visual stimulation.
Stack:
- Python
- OpenCV
- Machine learning
- MatLab