Guerbet has been granted a patent for a mechanism that enables seed relabeling for seed-based slice-wise lesion segmentation. The mechanism uses a lesion mask to generate a distance map and performs local maxima identification to select a set of local maxima. Seed relabeling is then performed based on the distance map and set of local maxima to generate a set of seeds, which are used for image segmentation to form a split lesion mask. GlobalData’s report on Guerbet gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on Guerbet, Nanorobotics was a key innovation area identified from patents. Guerbet's grant share as of September 2023 was 54%. Grant share is based on the ratio of number of grants to total number of patents.
Seed relabeling for slice-wise lesion segmentation in medical images
A recently granted patent (Publication Number: US11749401B2) describes a method for segmenting lesions in medical images using a data processing system. The method involves several steps to accurately identify and separate lesions in a three-dimensional medical image volume.
The first step of the method involves receiving a lesion mask, which corresponds to the detected lesions in the medical image volume. Each detected lesion has a lesion contour, and the method generates one or more lesion contours based on the received mask.
Next, a distance map is generated for a given two-dimensional slice in the medical image volume. This distance map represents the distance from each voxel in the slice to at least one lesion contour.
The method then performs local maxima identification to select a set of local maxima from the distance map. Each local maximum has a value greater than its immediate neighbor points and corresponds to a region comprising the immediate neighbor points.
Seed relabeling is performed based on the distance map and the set of local maxima to generate a set of seeds. Each seed represents the center of a distinct component of the corresponding lesion contour. The seed relabeling process involves merging the regions corresponding to at least two local maxima based on a predetermined relationship between them.
Finally, image segmentation is performed on the lesion mask based on the set of seeds to form a split lesion mask. This segmentation process helps to accurately separate and identify different lesions within the medical image volume.
The patent also includes claims for a computer program product and an apparatus that implement the described method. These claims cover the use of non-transitory computer-readable storage media and a processor to execute the steps of the method.
Overall, this patent describes a method for segmenting lesions in medical images that utilizes distance maps, local maxima identification, seed relabeling, and image segmentation techniques. The method aims to improve the accuracy and efficiency of lesion segmentation in three-dimensional medical image volumes.