• A
  • A
  • A
  • АБB
  • АБB
  • АБB
  • А
  • А
  • А
  • А
  • А
Обычная версия сайта
Глава в книге
ALOE: Boosting Large Language Model Fine-Tuning with Aggressive Loss-Based Elimination of Samples

Demidovskij A., Трутнев А. И., Тугарев А. М. et al.

In bk.: Frontiers in Artificial Intelligence and Applications: 27th European Conference on Artificial Intelligence, 19–24 October 2024, Santiago de Compostela, Spain. Vol. 392. IOS Press Ebooks, 2024. P. 3980-3986.

Препринт
DAREL: Data Reduction with Losses for Training Acceleration of Real and Hypercomplex Neural Networks

Demidovskij A., Трутнев А. И., Тугарев А. М. et al.

NeurIPS 2023 Workshop. ZmuLcqwzkl. OpenReview, 2023

Visual geometry and 3D image processing

2023/2024
Учебный год
ENG
Обучение ведется на английском языке
6
Кредиты

Course Syllabus

Abstract

This online course offers an introduction to the main tasks of 3D computer vision. We discuss both classical algorithms and modern neural network approaches. Students will get familiar with the theoretical description of the methods and obtain hands-on experience with practical tasks. Everyone who has been introduced to 2D computer vision and wants to extend their knowledge to the 3D world will be interested in this course.
Learning Objectives

Learning Objectives

  • The aim of this online course is to introduce students to 3D computer vision algorithms. Students will work with classic approaches and deep learning models for depth estimation and processing point clouds, understanding scene geometry for augmented reality and implementing vision perception for self-driving cars. Both core theory related to 3D computer vision and practical skills will be presented through the course in order to give students a solid background for using learned material in the future.
Expected Learning Outcomes

Expected Learning Outcomes

  • Distinguish the main tasks of 3D image processing
  • Distinguish the basics of 3d image processing
  • Apply epipolar geometry
  • Estimate depth based on various input modalities
  • Apply depth estimation models in practice
  • Distinguish applications of point clouds
  • Apply PointNet architecture for different tasks
  • Distinguish simultaneous localization and mapping (SLAM) methods
  • Apply SLAM in practice
  • Distinguish between approaches for multi-view generation
  • Apply NeRF algorithm for new view generation
  • Being able to perform 3D image processing
  • Apply neural network models in practice
Course Contents

Course Contents

  • Introduction to 3d image processing
  • Basics of 3D image processing
  • Depth Estimation
  • Point Clouds
  • Localization and mapping
  • Multi-View Generation
  • Final task with instructor’s evaluation
Assessment Elements

Assessment Elements

  • non-blocking Экзамен
  • non-blocking Практическая работа
Interim Assessment

Interim Assessment

  • 2023/2024 1st module
    0.5 * Практическая работа + 0.5 * Экзамен
Bibliography

Bibliography

Recommended Core Bibliography

  • Hartley, R., & Zisserman, A. (2015). Multiple View Geometry in Computer Vision (2nd ed). Australia, Australia/Oceania: Cambridge University Press. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsbas&AN=edsbas.1FEA5378
  • Mathematical methods in computer vision, , 2003
  • Richard Szeliski. (2010). Computer Vision: Algorithms and Applications. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsbas&AN=edsbas.E8FCD1BD

Recommended Additional Bibliography

  • Инженерная 3D - компьютерная графика : учебник и практикум для акад. бакалавриата, Хейфец, А. Л., 2015

Authors

  • Лабанина Алина Валерьевна
  • MAKAROV Ilia ANDREEVICH