First of All : Be sure that you are using GPU in colab.

🎯Source codes of the shoe sole detection project. I developed when I worked as an computer vision engineer at Eurobotik. Code Language is Pyhton. YOLO version is v8 (latest version during project development). I used Roboflow to pull the data sets. You can use your own dataset with api keys or use my datasets from roboflow(https://app.roboflow.com/yolov8-hlm3z/detectingshoesolesyolov8/3). I used version2 of my dataset in this repo but you can use whatever you want the main difference between second and third is that classes; v3 has only 1 classes which is soles despite v2 have 3 classes. The main idea about reducing class number is to be more specific for soles. Pay attention to the paths used in the project, update according to your own project status.

🎯With prediction.py you can use the live webcam application on any IDLE. Don't forget to add the best.py file that we previously created in Google Colab.

Angle.py = In this file, I took project one step further. In addition to the shoe sole, we can now estimate the angles.

Livecam.py = Alternative and modified live-time shoe sole angle and model finding. Correction of angle detection problems, %10 error solved.