Skip to main content
This page covers the full structure of the course: what is taught, when, and how you are assessed.
The course runs August–November 2025 at Pontificia Universidad Católica de Chile. There are 28 classes, meeting on Tuesdays and Thursdays. Professor: Domingo Mery.

​
Learning progression

The course is organized into five chapters. Each chapter builds on the previous one, moving from foundations to application to responsibility.

Cap00 — General lines

Course presentation, bibliography, and exam preparation. Establishes expectations and provides the reference materials used throughout the semester.

Cap01 — Introduction

Definitions of computer vision, motivating applications, and a two-part history of the field from early perspective geometry to the deep learning era.

Cap02 — Computational geometry

The mathematical backbone of the course. Homogeneous coordinates, 2D and 3D transformations, homographies, camera calibration, RANSAC, epipolar geometry, trifocal geometry, and 3D reconstruction.

Cap03 — Deep learning

Convolutional neural networks, object detection (YOLO), facial analysis, semantic segmentation (UNet), generative adversarial networks (GANs), anomaly detection, CLIP, Transformers, Visual Transformers (ViT), and diffusion models.

Cap04 — Ethics & AI

The good, the bad, and the ugly of AI. Facial recognition ethics, fairness, explainability (MinPlus / saliency maps), federated learning, data-protection law, and responsible deployment practices.

​
Full 28-class schedule

Exam support resources (Class 28)
History lectures cover the arc from Renaissance perspective machines through the deep learning revolution.
Practice exercises for this chapter — points/lines, transformations, and homographies — are available as PDFs in the course repository.
Assignment T02 is released in Class 18 and covers deep learning project work.
The essay assignment T03 is described on Canvas. The quiz E11 in Class 27 is also submitted via Canvas.

​
Grading and assignments

The course has three main assignments plus quizzes:
Past exam questions and solutions from 2023 and 2024 are available in the course repository under Cap00_Lineas_Generales/examenes/. Class 28 is dedicated entirely to exam preparation.

​
Tools and environment

The course uses two primary computing environments:
  • Python / Google Colab — the primary tool for all exercises and assignments. Notebooks run in the browser with free GPU access.
  • MATLAB — used for selected geometry and calibration demonstrations.
All Colab notebooks are linked from the class schedule in the course GitHub repository.