Computer Vision

Credits
6
Types
  • MIRI: Elective
  • MDS: Elective
  • MEI: Elective
Requirements
This subject has not requirements , but it has got previous capacities
Department
ESAII
Computer vision is a field within computer science focused on extracting meaningful information from images or sequences of images. Its range of applications is continuously expanding and includes facial recognition, early diagnosis of diseases, object and person detection and localization, gesture-based interaction with systems, robot navigation, and autonomous driving.

By the end of the course, students will be able to analyze, design, implement, and evaluate image analysis methods and techniques, meeting requirements related to response time, reliability, and cost/efficiency.

Teachers

Person in charge

Competences

Transversal Competences

Transversals

  • CTR6 - Capacity for critical, logical and mathematical reasoning. Capability to solve problems in their area of study. Capacity for abstraction: the capability to create and use models that reflect real situations. Capability to design and implement simple experiments, and analyze and interpret their results. Capacity for analysis, synthesis and evaluation.

    • Objectives

    1. Understand the limitations and capabilities of computer vision algorithms.
      Related competences: CTR6,

    Contents

    1. Fundamentals of digital imaging
      Types of images according to the different areas. Intensity images. Color images. 3D image for tomography, MRI, ultrasound images, etc. Color Spaces.
    2. Digital image processing
      Gray level transformations. Linear operators. Convolution. Image enhancement and smoothing. Contour detection. Nonlinear operators. Morphological filters. Geometric transformations.
    3. Image segmentation.
      Image binarization: global, local. Image segmentation: watershed, k-means, grouping by color.
    4. Image descriptors
      Numerical shape descriptors, regions, color histograms, Fourier descriptors, singular points, Haar.
    5. Image recognition using Machine Learning
      Image recognition and classification using descriptor vectors. Perceptual hash of images.
    6. Image recognition using Deep Learning
      Main deep neural networks for object detection and localization in images.

    Activities

    Activity Evaluation act


    Development of topic 1 of the subject



    Theory
    0h
    Problems
    4h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    8h

    Development of topic 2 of the subject



    Theory
    0h
    Problems
    8h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    12h

    Development of topic 3 of the subject



    Theory
    0h
    Problems
    4h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    8h

    Development of topic 4 of the subject



    Theory
    0h
    Problems
    8h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    12h

    Development of topic 5 of the subject



    Theory
    0h
    Problems
    4h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    8h

    Development of topic 6 of the subject



    Theory
    0h
    Problems
    4h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    8h

    Development of a real computer vision project


    Objectives: 1
    Theory
    0h
    Problems
    16h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    40h

    Presentation of the computer vision project



    Theory
    0h
    Problems
    4h
    Laboratory
    0h
    Guided learning
    0h
    Autonomous learning
    1.9h

    Teaching methodology

    The teaching methodology will generally be deductive in nature. An attempt will be made to avoid the expository/lecture method.
    The approach will be based on:
    - proposing a problem
    - trying to solve it
    - adding the necessary pieces of theory to be able to solve the problem adequately.

    During the practices, cooperative learning will also be worked on, for solving the problem as a team.

    Evaluation methodology

    The subject will be evaluated continuously. Throughout the course, a series of exercises will be requested that will serve to evaluate the student. There will be no final exam.

    The final grade for the subject (NF) will be obtained from the practices that are compulsorily done in class in person (LAB) and from the submissions of the practices that the student must work on at home (HW). Some exercises will be solved in groups and some individually. In group exercises the grade will be unique for all its components.
    The final grade will be calculated as follows:

    NF = Average(HW) * 0.5 + Average(LAB) * 0.5

    Where, HW and LAB represent the vector of grades for the work done at home and in the laboratory respectively.

    Bibliography

    Basic

    Previous capacities

    Basic statistics, elementary programming, algebra.