Difficulties of the project
The custom videos were of low quality, with poor lighting.
To improve the picture, we used four computer vision models. They predicted the position of an object in the frame, making the image clearer.
The neural network could color different frames in different ways; noises appeared in the video.
We have improved the work of the algorithm that was responsible for the clarity of the picture.
It was a time-consuming process: a 15- second video would take 15 minutes to process. This created a risk that the user would not wait or that the load on the serve would become too great
We indicated to the user exactly how much time was needed to queue a video to inform them of possible wait times. Added additional spare capacity which made provided stable operation