You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I believe the easiest way to use 360 degree images as data input would be to transform them into several (six would make it easiest but the number of input images can be changed) flat images, by following a preprocessing such as this.
Transforming the 360 image into several images allows them to go through the feature extraction network without having to retrain the whole ResNet + FPN.
If so , can you tell how , given that I use a 360 degree video as a data input ?
The text was updated successfully, but these errors were encountered: