Temporal and spatial correction for perfusion quantification system

Abstract

A medical imaging system includes an imaging device for producing temporally spaced image representations, a memory for storing the image representations, and a processor. The processor aligns the image representations by iteratively deriving a transform between progressively smaller sub-sets of pixels in temporally adjacent images until the sub-sets converge, and, applies the transform to the temporally adjacent images. The processor further sorts selected pixels in a region of interest spatially within each of the image representations for further processing by classifying pixels in the images according to a characteristic indicative of an absorption rate of the contrast agent. The processor further establishes correspondence between the selected pixels over successive image representations by generating masks for a particular classification of pixels, based on the generated masks, computing textural properties of the particular classification of pixels across the temporally spaced images, and computing co-occurrence matrices between successive image representations based on the textural properties. The processor further determines and displays a curve indicative the contrast agent in the selected pixels over time.