Automatic exposure correction of consumer photographs
ECCV'12 Proceedings of the 12th European conference on Computer Vision - Volume Part IV
Automated two-dimensional K-means clustering algorithm for unsupervised image segmentation
Computers and Electrical Engineering
| Hi-index | 0.43 |
A global image enhancement technique used to improve the visual quality of digitized photographs is described. The technique consists of two phases, viz., visual analysis and enhancement. In the visual analysis phase, approximate regions of an image that are visually significant are identified. Evaluation of visual significance is based on the features of focus, contrast, and texture which, in turn, are computed using some simple statistical measures. In the enhancement phase, a gray level transformation is applied to the entire image that pushes the average intensity of the visually significant regions towards neutral gray to provide the best detail in these regions. The transformation is obtained by simulating the exposure-density curve of photographic films and performing an exposure correction operation. Experimental results illustrating the operation and performance of the technique are presented. The proposed technique is useful for real-time applications as well as interactive enhancement of photographs