AI/ML: video stream tagging
Create a system to upload user videos for further analysis. The system should provide the ability to automatically analyze video. The system must be able to provide 10 concurrent user sessions and be linearly scalable to increase the number of concurrent users up to 500.
To solve the problem, modern State of the art machine learning models were used. The models are based on convolutional neural networks, which allows to speed up the learning process and forecasting by using GPU. Thus, if there are sufficiently powerful GPUs in the cluster, it is possible to achieve near real time prediction performance (the number of frames per second processed by the model approaches the video FPS value). The module allows to mark up video (photo) dataset in parallel, which is later used for model trainig.
Developed a multi-node linearly scalable platform for automatic video analysis using hardware GPUs, which made it possible to significantly increase the performance per cluster node. A user interface was developed for maintaining a video library, viewing and analyzing it, as well as allowing you to upload video and select analysis models for processing it when uploading.
Would you like to see
the full case study?
Fill out the form and we will
contact you right away