• new: Optional module arivis AI toolkit: 
• new: Deep Learning Trainer panel to setup and train a DL model 
• new: use the Draw Tool to draw segments to annotate regions for classes 
• new: utilizing of sparse annotations to overcome the need to draw on every pixel 
• new: drawn objects are sorted by class order - pixels are assigned to the topmost one (no need to cut out overlaps) 
• new: add and reuse existing segments for class definition 
• new: automatic clustering based on annotated backgound possible 
• new: run and monitor the DL training process right from Vision4D 
• new: directly open the trained DL model in the Analysis Pipeline for inference 
• new: export the trained DL model as ONNX 
• new: additional installer package to setup a whole DL Python environment