I realized that when annotating images loaded from the pc is not in the numbering sequence…
For example, img.1.jpg to img-n on the computer is not the same sequence on Roboflow.
Why this is a concern? When annotating the images ROBOFLOW mixes the images, therefore image37.jpg can be set as the first image for annotating, after that image25.jpg, then img209.jpg, and so on…
Is there an annotation method that keeps the image numbering sequence that it was in when loaded from the computer?
or have I loaded the images incorrectly as per Roboflow?
Another thing I noticed is that there isn’t a feature to detect the ROI’s for objects, especially for activity detection as the time series alters as per the vector flows of such.
There is a feature to copy the ROI from the previous image but that is useless as well as pointless if the object of interest is continuously moving across the different frames.
Why? I am constantly redrawing the ROI’s for the objects and I have thousands of images to process with multiple labels (extremely painful).
MY request: is there an easier method that I must have missed somewhere?
Maybe a feature that tracks or detects the chosen ROI in the images whether it moves or not?
Can someone guide me here, please? This is extremely painful and a tedious process even if the best song is running in my pods in the background.