The simplest example of features that perform image enhancement (a sort of transformation) is brightness / contrast control. But for sure, there is a good potential for additional real time image / video / 3d enhance enhancement apps.
Below I will provide several examples.
(1) Rotate / flip images in Labeling Interface. It might happen that some images are already imported but the image is “mistakenly” rotated. This app might provide easy correction
(2) Crop images in Labeling Interface. Suppose that we are interested in a very particular area within an image, so it might be convenient to crop this image in real time
(3) Quick way to anonymize a part of an image. Suppose that we noticed that an image contains a confidential or other information that we would not like to share. So it might be convenient to anonymize some area within an image in real time
(4) 3D labeling. App to automatically hide the “ground” in 3D point cloud labeling interface. It might be easy to label cars when “irrelevant” objects are hidden
(5) Radiology. I hope someone knows the domain and could suggest something useful
(6) Artificial images. That might sound a bit crazy, but suppose the following. Suppose we have a model that can generate objects of interest, say persons. So, in the Labeling Interface, we select an area with a rectangle, the model produces a picture of a person and, as a result, we have a new image that contains “generated person”. As a side note, in our consultancy practise, we used a similar approach to artificially generate labeled cracks on the roads.
If you have more ideas that fit into the list above, please, share them in the comments.