The Image Input and Image Template tools in the Computer Vision category do not accept relative path names. This makes it difficult to write a workflow in a test environment and then run it in the operating environment. (Someone in the operating environment has to diddle with the Image Input path name, and the Image Template seems to require a complete re-training to spot the areas needed for scanning.)
ALSO the drop-down choices for "the most accurate option for what product component this enhancement would effect (sic)" does not offer the correct option of "Computer Vision".