This repository contains practical exercises completed during the course Vision and Cognitive Systems at UniPd. The class teaches the concepts, methods, and technologies underlying computer vision and cognitive systems, including modern cognitive services such as APIs and cloud services that assist developers in building artificial intelligence applications.
-
Linear Regression: Machine Learning, Linear Regression
-
Image Filtering: Image Filtering, Derivative Filters, Edge Detection
-
SIFT: Local Visual Features, SIFT Algorithm, Feature Matching
- Utilized OpenCV's SIFT algorithm for detecting keypoints and descriptors of an image.
- For feature matching, used Brute-Force (BF) Matcher, K-NN matcher (in order to apply the ratio test as reported by D. Lowe), and FLANN-based matcher.
-
Bag-of-visual-words for Image Classification: BoW (Bag of visual words) for image classification
- Steps:
- Image features quantization
- Histogram of visual words
- Classification (NN/k-NN/SVM)
- Steps:
-
Multilayer Perceptron (MLP): Multilayer Perceptron using PyTorch
- Classifying CIFAR10 images using Multilayer Perceptrons (MLPs).
- Did data normalization, hyper-parameters fine-tuning, data augmentation, and weights initialization.
-
CNNs: CNNs for image classification using PyTorch