We propose a new multiframe algorithm to enhance the spatial resolution of frames in video sequences. Our technique specifically accounts for the possibility that motion estimation will be inaccurate and compensates for these inaccuracies. Experiments show that our multiframe enhancement algorithm yields perceptibly sharper enhanced images with significant signal-to-noise ratio (SNR) improvement over bilinear and cubic B-spline interpolation.

No abstract

No abstract

No abstract

The Mumford-Shah model is extensively used in image segmentation. Its energy functional causes the content of the segments to remain homogeneous and the segment boundaries to become short. However, the problem is that optimization of the functional can be very slow. To attack this problem, we propose a reduced two-phase Mumford-Shah model to segment images having one prominent object. First, initial segmentation is obtained by the k-means clustering technique, further minimizing the Mumford-Shah functional by the Douglas-Rachford algorithm. Evaluation of segmentations with various error metrics shows that 70 percent of the segmentations keep the error values below 50%. Compared to the level set method to solve the Chan-Vese model, our algorithm is significantly faster. At the same time, it gives almost the same or better segmentation results. When compared to the recent k-means variant, it also gives much better segmentation with convex boundaries. The proposed algorithm balances well between time and quality of the segmentation. A crucial step in the design of machine vision systems is the extraction of discriminant features from the images, which is based on low-level segmentation which can be obtained by our approach.