A multiresolution texture segmentation approach with application to diagnostic ultrasound images

A multiresolution texture segmentation approach is presented which addresses characterization, resolution, and time considerations. This approach uses simulated annealing to minimize an energy function based on the resolution of different texture blocks in an image. A rigorous experimental procedure is developed to ensure that accuracy and speed criteria are met. Semireal images, created by sampling a series of diagnostic ultrasound images of an ovary in vitro, are tested to produce statistical measures on the performance of the approach. The ultrasound images themselves are then segmented to determine whether the approach can achieve accurate results for the intended ultrasound application. Experimental results confirm that the multiresolution approach converges faster and produces better segmentation results than the single-level approach.<>