Unusual Image Processing Solutions
I found one of those things that you want to talk about once you saw them, because they are just too cool. Ok, probably I’m exaggerating now but I think it’s cool anyway.
I wanted to talk about Image Processing applications, cool ones, that don’t require a lot of specific knowledge to understand them, so people whom don’t know about the field could start paying attention to it. Finally, today I end up with an idea that is very interesting: maze solve through Image Processing!
Loren and Cris present a good analysis on how to perform the task, using MATLAB and other tools. In general, the solution is based on operations in the image that separate the connected walls, treating them as regions the path to get out lies on the boundaries of those regions. I won’t give details on how to do it, go to those sites and read the entire solution. Instead, I want to point out how interesting it is to found solutions in fields one won’t imagine.
These guys are doing an interesting job too. They create mazes from images using Image Processing algorithms with a few inputs from user. So now, computers can solve and create puzzles using images. Just like us!
Image Processing offers a lot of solutions for a lot of different problems we are not aware of yet. We use our eyes to process a lot of information in our lives. This leads the question: How many problems machines will be able to solve when they perceive the world as we do? 💡
We need to think out-of-the-box and go for a solution that is different from the common one. We need to see the problem from a different angle and tackle it from there. This approach give us a way to step forward in some fields that are stuck; researchers solve problems in the same way again and again, just tweaking ideas and doing small changes and trying new version of the same idea, without a major contribution. I think that this behavior continues because it is easy to get accepted by peer review, just because is related to the trend. If this behavior continues, however, science develop will slow down.