Signal Shape Modeling

About

Speckle noise is the most common type of noise, which degrades image appearance as well as clinical interpretation accuracy/reliability, with various medical imaging modalities, including optical coherence tomography (OCT) and ultrasound. Although there have been many attempts to reduce speckle noise using various image processing techniques, their heavy computational load and/or image quality degradation (blurring effect) prevented them to be practical solutions.   Technology: Investigators have developed a novel method overcoming these limitations. By modeling the meaningful signal shape, speckle noise can be selectively and intelligently suppressed.    Advantages: This invention first applies conventional smoothing filtering to the entire image, but the outcome will not become the end result. Instead, the filtered image represents a model of meaningful signal. Then the invention searches through the image for significant outliers having high reflectivity (one of the major trait of speckle noise) and suppress them to the level of the afore mentioned model. This two phase approach has never been described. Since the only alteration to the original image will be selective speckle noise reduction, this generally does not affect the image quality, which cannot be achieved by conventional filtering techniques.