Academic projects for a Computer Vision course. Each lab covers a distinct topic in image processing or deep learning, implemented in Python as Google Colab notebooks.
Introduction to OpenCV and Matplotlib. Covers loading images, creating copies, drawing shapes, and adding text annotations.
Applies and compares classical image filters — median, mean, Gaussian blur, and erosion — followed by edge detection using the Laplacian operator.
Builds and trains a fully connected neural network on the MNIST handwritten digit dataset using PyTorch. Covers data normalization, one-hot encoding, nn.Sequential, Adam optimizer, MSE loss, and training loop with history visualization.
Real-time face detection using the MediaPipe Tasks API and the pre-trained BlazeFace (short-range) model. Draws bounding boxes and facial keypoints on detected faces.
Trains a convolutional neural network (Conv2D + MaxPooling + Dropout) on the CIFAR-10 dataset (10 object classes) using Keras. Includes model saving and inference on custom images.
Classifies clothing items from the Fashion-MNIST dataset (10 categories) using a dense neural network in Keras/TensorFlow. Includes evaluation, prediction visualization, and inference on custom images.
| Area | Tools |
|---|---|
| Image processing | OpenCV, NumPy |
| Deep learning | PyTorch, TensorFlow / Keras |
| Computer vision | MediaPipe (BlazeFace) |
| Visualization | Matplotlib |
| Environment | Python 3, Google Colab |