Optical Enquiry
Member
- Messages
- 19
- Likes
- 0
I have studied all the images that the professor provided and all the images on pages that he linked to and in every case the result was just bad. It becomes even more noticeable when I push the contrast to the extremes in LAB mode of both original and filtered images. In that context I find it odd that we're being asked to find a way to achieve those same (bad) results. That's the problem I have with the whole idea; I haven't seen any benefit at all. I'm a visual guy, I've developed a keen eye for these things and I just want to see an image that visually got better. Even going by the actual numbers, I still see no improvement at all, only images that have their quality degraded.
I agree, our results are not always pleasing. However they were made with an engineering community in mind and their main point is to demonstrate the technology. More subtle changes would not have made the point as clearly (to engineers). The big issue is explain to computer vision people the difference between hue and color, which is a thing that most artists are familiar with.
We are thinking now of ways to now get this approach into the hands of people that can use it and can produce good looking results. If you could go to post number 10 and look at the image. Could you find a use for the tool described in the schematic?
Last edited: