Title :
Image Analysis Under Varying Illumination
Author :
Zeng, Hengli ; Trussell, H.J.
Author_Institution :
Dept. of Electr. & Comput. Eng., North Carolina Univ., Raleigh, NC, USA
Abstract :
Often in dealing with images, the training data must be extracted for a limited data set. In particular, the illumination conditions of the sample images is limited and, in many cases, unknown. In this paper, we show that artificial variation of the illuminant of hyperspectral images can be used to overcome the limitations of a small sample set.
Keywords :
image sampling; learning (artificial intelligence); lighting; neural nets; hyperspectral image; image analysis; sample image; sample set; training data; varying illumination; Artificial neural networks; Hyperspectral imaging; Image analysis; Lighting; Neural networks; Neurons; Object detection; Pixel; Reflectivity; Vectors; Image processing; Lighting control; Neural network applications; Pattern classification;
Conference_Titel :
Image Processing, 2006 IEEE International Conference on
Conference_Location :
Atlanta, GA
Print_ISBN :
1-4244-0480-0
DOI :
10.1109/ICIP.2006.312625
