Electrical_Experimental_Course

This repository contains various Python scripts and exercises covered in the Electrical Experimental Course.

Project Structure

Electrical_Experimental_Course/
│
├── python/
|   ├── mid_class/
│       ├── calculatorAPI.py            # Simple calculator API
│       ├── class3_for.py               # For-loop example
│       ├── class3_if.py                # If-else condition example
│       ├── class3_while.py             # While-loop example
│       ├── class4_class.py             # Python classes example
│       ├── class4_def.py               # Function definitions
│       ├── class4_import_practice.py   # Importing custom modules
│       ├── class4_import_sqrttest.py   # Using math.sqrt example
│       ├── class5_match.py             # Match-case example
│       ├── class5_raise.py             # Raising exceptions
│       ├── class5_try.py               # Try-except for error handling
│       ├── class6_break_ex1.py         # Break statement usage
│       ├── class6_continue_ex1.py      # Continue statement usage
│       ├── file1.txt                   # Test file 1
│       ├── file2.txt                   # Test file 2
│       ├── file3.txt                   # Test file 3
│      
|   ├── yolov8_python/
│       ├── yolov8_test.py              # YOLOv8 model testing script
│       ├── yolov8_train.py             # YOLOv8 model training script
│       ├── yolov8_train.py             # YOLOv8 model validing script
│
├── camera/
│   ├── hand_test.py                # Automatically predicts hand gestures and displays

YOLOv8 Model Description

This project includes a YOLOv8 model trained for hand gesture recognition. Below is the detailed explanation of the dataset, training process, and results:

Dataset

• Total Images: 600 (original)
  • Labels: Five hand gestures (skr, gun, spiderman, rock, italy) and one additional class other.
• Data Augmentation:
  • Techniques: Image flipping and mosaic augmentation.
  • Final Dataset Size: 909 images.
  • Training set: 729 images (80%).
  • Validation set: 120 images (13%).
  • Test set: 60 images (7%).

Training

• Epochs: 100 iterations.
• Model: YOLOv8.
• Framework: Ultralytics.

Validation Results

The following are the validation results obtained after training:

Class	Images	Instances	Precision (P)	Recall (R)	mAP50	mAP50-95
All	120	100	0.967	0.971	0.979	0.966
Gun	120	21	0.997	0.952	0.965	0.930
Rock	120	19	0.945	1.000	0.993	0.993
Gun	120	20	0.946	0.950	0.957	0.952
Skr	120	20	0.996	1.000	0.995	0.995
Spiderman	120	20	0.950	0.954	0.983	0.962

Results

The final prediction video can be found here:

results.mp4