Perspective Transformation Layer
Document Type
Conference Proceeding
Publication Date
1-1-2022
Abstract
Incorporating geometric transformations that reflect the relative position changes between an observer and an object into computer vision and deep learning models has attracted much attention in recent years. However, the existing proposals mainly focus on the affine transformation that is insufficient to reflect such geometric position changes. Furthermore, current solutions often apply a neural network module to learn a single transformation matrix, which not only ignores the importance of multi-view analysis but also includes extra training parameters from the module apart from the transformation matrix parameters that increase the model complexity. In this paper, a perspective transformation layer is proposed in the context of deep learning. The proposed layer can learn homography, therefore reflecting the geometric positions between observers and objects. In addition, by directly training its transformation matrices, a single proposed layer can learn an adjustable number of multiple viewpoints without considering module parameters. The experiments and evaluations confirm the superiority of the proposed layer.
Identifier
85171996523 (Scopus)
ISBN
[9798350320282]
Publication Title
Proceedings 2022 International Conference on Computational Science and Computational Intelligence Csci 2022
External Full Text Location
https://doi.org/10.1109/CSCI58124.2022.00250
First Page
1395
Last Page
1401
Grant
CIF/SaTC-2104267
Fund Ref
National Science Foundation
Recommended Citation
Khatri, Nishan; Dasgupta, Agnibh; Shen, Yucong; Zhong, Xin; and Shih, Frank Y., "Perspective Transformation Layer" (2022). Faculty Publications. 3366.
https://digitalcommons.njit.edu/fac_pubs/3366