Multiresolution Approach to Processing Images for Different Applications

This book presents theoretical and practical aspects of the interaction between low and high level image processing. Multiresolution analysis owes its popularity mostly to wavelets and is widely used in a variety of applications. Low level image processing is important for the performance of many high level applications. The book includes examples from different research fields, i.e. video surveillance ; biomedical applications (EMG and X-ray) ; improved communication, namely teleoperation, telemedicine, animation, augmented/virtual reality and robot vision ; monitoring of the condition of ship systems and image quality control. Nowadays computers are expected to perform complex tasks involving the processing of huge amounts of data. People are often unaware that such intensive operations are required in computer vision tasks. Different phenomena related to pixel/voxel or global size can cause failure of higher vision applications. Such phenomena include illumination variations, noise, camera jitter, shadows, visibility, weather conditions, etc. This paper analyzes the influence of lower vi- sion tasks (e.g. denoising or thresholding) on higher vision tasks (e.g. motion segmentation or product quality). The interaction between lower and higher vision is illustrated with examples of visual quality control and advanced visualization in marine communications used to decrease the stress felt by seafarers due to their separation from their families.