Share Email Print
cover

Proceedings Paper

Robust CMOS camera module lens calibration by support vector machine regression
Author(s): Chan-Yun Yang; Gene Eu Jan; Yung-Yuan Chen
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Motivated by the powerful computational capability in the emerging hardware, an applicable paradigm with its embedded lens calibrator is proposed. The proposed new paradigm for the relationship between the image provider and the image processor shows both the functional and economical merits. The paper first focuses on the developed of the embedded lens calibrator. An underlying support vector machine base regression (SVR) is hence employed as the key to achieve the goal. Based on the structural risk minimization, the SVR, employed as the calibration regressor, simultaneously minimize both the model complexity and empirical error, and create an estimator with a wide margin. The wide margin in regression represents a smooth approximation function for the lens calibration in which variances commonly existed in the CMOS camera modules can tolerably be eliminated. The variance tolerability achieves the calibration function a high robustness, and would conduct potentially the success of the proposed paradigm.

Paper Details

Date Published: 31 January 2013
PDF: 10 pages
Proc. SPIE 8759, Eighth International Symposium on Precision Engineering Measurement and Instrumentation, 875949 (31 January 2013); doi: 10.1117/12.2014433
Show Author Affiliations
Chan-Yun Yang, National Taipei Univ. (Taiwan)
Gene Eu Jan, National Taipei Univ. (Taiwan)
Yung-Yuan Chen, National Taipei Univ. (Taiwan)


Published in SPIE Proceedings Vol. 8759:
Eighth International Symposium on Precision Engineering Measurement and Instrumentation
Jie Lin, Editor(s)

© SPIE. Terms of Use
Back to Top