Dog Breed Classifier
AI-powered dog breed classification using machine learning in Python.
Tech Stack
Overview
An artificial intelligence project built in Python for classifying dog breeds from images. The project was developed as part of the ORI (Osnove Računarske Inteligencije — Foundations of Computational Intelligence) course at the Faculty of Technical Sciences. It uses machine learning techniques to train and evaluate a model capable of identifying dog breeds from visual input.
01 — The Problem
Applying machine learning concepts from an AI course to a real classification task — training a model to distinguish between dog breeds based on image data, without relying on pre-built cloud services.
02 — The Solution
Built a Python-based ML pipeline that preprocesses image data, trains a classification model, and evaluates its accuracy across multiple dog breeds. The entire pipeline — from data loading to evaluation — is implemented from fundamentals as part of the coursework requirements.
03 — The Outcome
A working image classification system that demonstrates practical application of AI/ML fundamentals, including model training, evaluation metrics, and image preprocessing techniques.