How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
Even with all the advancements being made in imaging, Euresys’ Damhaut says there’s more possible. “I think that machine vision software still has a long way to go. Many of the algorithms and application software used today have to be tweaked via numerous parameters to perform correctly in a given application. This puts them out of reach of numerous inexperienced users. Take for example an application involving reading characters. Current OCR (optical character recognition) algorithms still perform poorly compared to the human brain in terms of recognition rate. I see a lot of potential improvement coming in the future addressing the robustness and user friendliness of libraries and applications.”
This month Euresys will launch a new Open eVision release (1.1) that includes several advances, says Muriel Ancion, VP of Communications. “The key feature is the introduction of Flexible Masks that add flexibility to the shape of the ROIs,” she explains. “A mask represents a two-class segmentation of pixels that separates the associated image in ‘do-care’ areas (that must be considered for the processing) and ‘don’t care’ areas (that should not be considered). The masks support complex and disconnected shapes.”