OpenCV Python Course — Automated Waste Detection and Classification Using OpenCV and Convolutional Neural Network (CNN)
Below is a syllabus for 20 sessions (1.5 hours each) for a thesis titled:
Automated Waste Detection and Classification Using OpenCV and Convolutional Neural Network (CNN)
Stage 1: Fundamentals & Dataset Preparation (Sessions 1–5)
Session 1:
-
Introduction to computer vision and its role in waste management
-
Literature review of image-based waste classification systems
Session 2:
-
Installation of Python, OpenCV, TensorFlow/Keras, and Jupyter Notebook
-
Introduction to basic image processing using OpenCV
Session 3:
-
Introduction to waste types: organic, inorganic, hazardous
-
Analysis of open-source datasets (such as TrashNet or Roboflow datasets)
Session 4:
-
Image preprocessing: resize, grayscale, histogram equalization
-
Data augmentation techniques to improve model generalization
Session 5:
-
Introduction to CNN: basic architecture (Conv, Pool, Flatten, Dense)
-
Implementation of simple CNN with Keras
Stage 2: Model Training & Evaluation (Sessions 6–12)
Session 6:
-
Training CNN model for classification of 3–5 waste types
-
Evaluation metrics: accuracy, loss, confusion matrix
Session 7:
-
Fine-tuning model using pretrained MobileNetV2/ResNet
-
Transfer learning & performance comparison
Session 8:
-
Visualization of prediction results with Matplotlib
-
Analysis of classification errors
Session 9:
-
Dataset optimization and class balancing (if dataset is imbalanced)
-
Model re-evaluation with additional validation data
Session 10:
