Challenges and trade-offs for in-browser computer vision applications

For client applications, computer vision technology claims to be an excellent way to improve interaction and increase the interactivity of the user interface. The technology has proven itself well for native applications and is going on the web.

However, engineers need to be prepared for the fact that the web browser is still an extremely immature platform for such tasks. If one day you decide to use computer vision in a web application, you are likely to face several problems and limitations, which will be discussed in this talk.