Conference 12100 > Paper 12100-38
Paper 12100-38

3D shape object reconstruction with non-Lambertian surface from multiple views based on deep learning

5 April 2022 • 6:00 PM - 8:00 PM EDT

Abstract

This paper proposes an algorithm to estimate a surface reflectance model of 3-D shape and parameters from multiple views. To solve the problem of determining the views without Lambert's surface, a 3-D reconstruction algorithm based on a convolutional neural network is proposed. To train the neural network, we apply two stages. At the first stage, the encoder is trained for the descriptor description of the input image. In the second step, a fully connected neural network is added to the encoder for regression for choosing the best views. The coder is trained using the generative adversarial methodology to construct a descriptor description that stores spatial information and information about the optical properties of surfaces located in different areas of the image. The codec network is trained to recover the defect map (depends directly on the sensor and scene properties) from a color image. Experimental results on both synthetic and real objects are given.

Presenter

Moscow State Univ. of Technology "STANKIN" (Russian Federation)
Evgenii Semenishchev is the research of the Center for Cognitive Technology and Machine Vision at Moscow State University of Technology “STANKIN,” Moscow, Russian Federation. He received his BS (2005), MS (2007) in the communication system from the South-Russian State University of Economics and Service, and his Ph.D. in technics from Southern Federal University (2009). Evgenii is a member of SPIE society. His research interests include image processing, image fusion and computer vision.
Author
Moscow State Univ. of Technology "STANKIN" (Russian Federation)
Author
The City Univ. of New York (United States)
Presenter/Author
Moscow State Univ. of Technology "STANKIN" (Russian Federation)
Author
Alexandr Zelensky
Moscow State Univ. of Technology "STANKIN" (Russian Federation)
Author
The City Univ. of New York (United States)