Our software performs image operations (for example rotation, deskew etc) and we would like to create tests for those. However the operations are not guaranteed to always produce the same result at the byte level. Some noise here and there, some artifacts or dithering might change the result image in a way not apparent to the human eye but breaking a per byte comparison. What would be the best way to test the results of such an operation (assuming I have a reference 'expected image')?
Aucun commentaire:
Enregistrer un commentaire