Doubt in Output Labels Of image Segmentation with U-net

I recently come across this image

in assignment we only extracted the max value out of all three dimensions from the masked image.So according to this for each pixcel the possible values are [0,70,100,90,220,153,234,128,244,107,142,156,220,180,81,150,230,180,250,110,170,150,170] but when i write a python code to extract all unique possible pixcel values present in the mask image after reducing max then i get this [0,1,2,3,4,5,6,…22] means from 0-22 why is it so i cannot get

The values you are seeing are the “labels”. That is what the mask pixels contain: the label for that corresponding pixel in the image. It is one value per pixel. If you wanted to render the mask as an image, you could then map those to the RGB color values from that table that you showed. Where is that from and are you sure that is connected to the dataset we have here?

yes the table is given in official page of CARLA car dataset Sensors reference - CARLA Simulator the link