For me it is interesting when people tells me "AI is very difficult" or "I couldn't do what you do". Where does the idea that AI is some sort of new rocket science?
First of all, Artificial Intelligence (AI) is a very big field. AI is not all about building self-driving cars or building complex algorithms - it is much more than that (and MUCH LESS as well).
AI teams are not only sitting down and building the next cutting-edge technology. They are also providing open source tech available to anyone. Yes, free for you to use - with very little amount of knowledge about the field. Because that is thee beauty about AI and being an AI practicioner. We want to educate people on how to use pree-existing tools to make a better world around us.
Think about it. You didn't invent Microsoft Word or Excel or PowerPoint, right? You probably don't even know how the technology works behind - but you still make use of it, right? SAME HERE. It is a tool for you to use, and the full AI product comes into place one someone has an interesting idea that can easily develop with tools outthere. Just like building an excel sheet that will estimate your companys' sells for the next month.
I show case in this video an idea I had a couple of months ago. There are two very amazing technologies out there for Computer Vision:
- 3D Photography using Context.aware Layered Depth Inpainting: https://github.com/vt-vl-lab/3d-photo-inpainting
- You Only Look Once (YOLO): https://github.com/AlexeyAB/darknet
The idea is simple. I wanted to see if I could improve YOLO's algorithms on occluded objects by first preprocessing the image through the first algorithm. This is AI. This is a product. This is new technology. This is what we do at work.
By using the 3d-photo-inpainting method we can brign an extra dimension to our picture, which will eventually help yolo to better detect "occluded" objects and improve the accuracy of the detection (for some of the frames).
The results are obviously only visual - but it does seem to make an improvement. I only needed basic knowledge of setting up the projects together, some programming background but I didn't use any of my previous AI knowldege for this.
What do you think?
GitHub project: https://github.com/coding-ai/yolt