Removal of Moving Objects and Inconsistencies in Color Tone for an Omnidirectional Image Database
This paper proposes a method for removing image inconsistencies which occur by an existence of moving objects or a change of illumination condition when an omnidirectional image database is generated. The database is used for archiving an outdoor scene in wide areas or generating novel view images with an image-based rendering approach. In related work, it is difficult to remove moving objects in an outdoor environment where illumination condition drastically changes, and to remove inconsistencies of color tone of images which included moving objects. The proposed method iterates the two processes which are the estimation of candidate region of moving objects and the achievement of color consistency to split regions. The color consistency is achieved by estimating linear color transformation parameters which change a histogram of an input image to that of the standard image.
KeywordsInput Image Candidate Region Standard Image Illumination Condition Outdoor Environment
Unable to display preview. Download preview PDF.
- 2.Sato, R., Ono, S., Kawasaki, H., Ikeuchi, K.: Real-time image-based rendering system for virtual city based on image compression technique and eigen texture method. In: International Conference on Pattern Recognition (2008)Google Scholar
- 4.Moroney, N.: Local color correction using non-linear masking. Color Science and Engineering Systems, Technologies, Applications, 108–111 (2000)Google Scholar
- 5.Tsuchida, M., Kawanishi, T., Murase, H., Takagi, S.: Sequential monte-carlo estimation of background image for background subtraction under changing illumination. In: Int. Conf. on Visualization, Imaging, and Image Processing, pp. 421–425 (2003)Google Scholar