Generating 3D Model From Video

Provided by: Academy & Industry Research Collaboration Center
Topic: Software
Format: PDF
In this paper, the process of 3D modeling from video is presented. Analyzed previous research related to this process, and specifically described algorithms for detecting and matching key points. The authors described their advantages and disadvantages, and made a critical analysis of algorithms. In this paper, the three detectors (SUSAN, Plessey and Forstner) are tested and compare. They used video taken with hand held camera of a cube and compare these detectors on it (taking into account their parameters of accuracy and repeatability). In conclusion, they practically made 3D model of the cube from video used these detectors in the first step of the process and three algorithms (RANSAC, MSAC and MLESAC) for matching data.

Find By Topic