Three-Dimensional Recovery from Image

Please enable javascript in your browser.

Karimi, Fatemeh | 2021

614 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 54386 (02)
University: Sharif University of Technology
Department: Mathematical Sciences
Advisor(s): Moghaddasi, Reza
Abstract:
3D reconstruction of digital images is one of the sub- branches of computer vision which is used in various fields such as computer graphics, compute animation, etc. in general, two approaches exist for solving 3D structure problem. In the first approach to obtain the model, we need a certain distance of the whole object level at a certain point and to do so, a sensor is used which is directly interacting with the object. For instance, we use lasers and estimate the depth using that depth to make 3 D model. In the second approach we use the images we have at different instaces of an object (sensor is the camera that is sensitive to light). Using these data and the machine learning algorithms we obtain depth that helps to the three - dimensional structure of the object. The advantage of this approach over the first approach is that it costs less. in this thesis, the goal is to recover 3D images and obtain the depth by using machine learning algorithms
Keywords:
Neural Network ; Computer Vision ; Machine Learning ; Three Dimensional Reconstruction ; Depth Estimation ; Image Reconstruction