Previous article https://qiita.com/asmg07/items/e3be94a3e0f0195c383b Last time I wrote about system introduction, so next I will write about learning.
By the way, I didn't write or decide the goal of the program, so for the time being I will set a goal. For the time being, I want to improve the program I made in my old class as a student. I'm writing. That's the reason for the development. For the time being, I like Nogizaka as a goal, so I wish I could make a Nogizaka personal identification classifier. It was a digression.
Aside from the aside, for the time being, I want to learn how to use the masks that various people are doing.
I think I'll use the image of whether it's done or not. I thought I'd say, but since it's a big deal, I'm going to use another model.
I would like to try using a car model. </ S>
I will make a model myself because it is a big deal.
I want to see Nogizaka-chan and others w
Step 1 Annotation
(1) Image creation
The following URLs are posted for the software to be used.
https://github.com/Microsoft/VoTT/releases
We will create the following data set.
I make it like this.
(2) Change to the dataset for Yolo
Well, I do various things, but the reference site will be helpful.
At the end, it looks like this.
I refer to the following site.
https://konchangakita.hatenablog.com/entry/2020/08/24/220000
(3) Learning at Yolo
Regarding learning, we are currently considering whether to do it in the local environment or use the Google one.
Recommended Posts