# US 12,002,215 B2 · Method and apparatus for compensating image segmentation lines > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,002,215 B2 - Document ID: US12002215B2 - Status: Granted - Kind: B2 - Application number: 17/506,637 - Filed: 2021-10-20 - Published: 2023-03-09 - Granted: 2024-06-04 - Earliest claimed: 2021-09-09 - Assignees: Acer Medical Inc. - Applicants: Acer Medical Inc. - Inventors: Shih-Ho Huang; Ming-Tzuo Yin - Web page: https://patentgrep.com/patent/US12002215B2 ## CPC - G06T 7/12: Edge-based segmentation - G06T 7/13: Edge detection - G06T 7/0012: {Biomedical image inspection} - G06T 7/181: involving edge growing; involving edge linking - G06V 10/26: Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion - G06V 10/82: using neural networks - G06T 2207/10101: Optical tomography; Optical coherence tomography [OCT] - G06T 2207/20081: Training; Learning - G06T 2207/20084: Artificial neural networks [ANN] - G06T 2207/30041: Eye; Retina; Ophthalmic - G06V 2201/03: Recognition of patterns in medical or anatomical images - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract A method and an apparatus for compensating image segmentation lines are provided. In the method, image segmentation is applied to a medical image captured by using a segmentation model to obtain a segmentation image including at least one segmentation line between multiple layers in the medical image. Convolution computation is then performed on the segmentation image by using a kernel of a trained classification model to predict a location of a next pixel connected to a current pixel in the respective segmentation line within the segmentation image, in which the pixel to be predicted is limited to a neighboring pixel of the current pixel in a prediction direction. The predicted pixels are connected to form a compensated segmentation line for each segmentation line. ## Claims ### Claim 1 A method for compensating image segmentation lines suitable for electronic apparatus with a processor, the method comprising: segmenting a captured medical image by adopting a segmentation model to obtain a segmentation image comprising at least one segmentation line between a plurality of layers in the medical image; performing a convolution operation on the segmentation image by adopting a kernel of a trained classification model to predict a position of a next pixel connected to a current pixel of the respective segmentation line in the segmentation image, wherein the predicted pixel is a neighboring pixel of the current pixel in a prediction direction; and connecting a plurality of predicted pixels to form a compensated segmentation line for the respective segmentation line. ### Claim 2 (depends on 1) The method for compensating image segmentation lines according to claim 1, further comprising: detecting a gap section in the respective segmentation line, and determining a kernel size of the classification model based on a size of the gap section. ### Claim 3 (depends on 2) The method for compensating image segmentation lines according to claim 2, wherein determining the size of the kernel of the classification model based on the size of the gap section comprises: calculating a distance between two endpoint pixels of the detected gap section in the prediction direction of the kernel, and setting a length of the kernel in the prediction direction accordingly, so that the length is greater than the distance. ### Claim 4 (depends on 1) The method for compensating image segmentation lines according to claim 1, further comprising: according to use requirements of the medical image, determining one or a combination of a kernel size and a padding size adopted by the classification model. ### Claim 5 (depends on 1) The method for compensating image segmentation lines according to claim 1, wherein the neighboring pixel comprises a next pixel of the current pixel in the prediction direction and neighboring pixels in a vertical direction perpendicular to the prediction direction of the next pixel. ### Claim 6 An apparatus for compensating image segmentation lines, comprising: an image capturing device, adapted to capture a medical image; a storage device, adapted to store a computer program; and a processor, coupled to the image capturing device and the storage device, and configured to load and execute the computer program to: segment the captured medical image by adopting a segmentation model to obtain a segmentation image comprising at least one segmentation line between a plurality of layers in the medical image; perform a convolution operation on the segmentation image by adopting a kernel of a trained classification model to predict a position of a next pixel connected to a current pixel of the respective segmentation line in the segmentation image, wherein the predicted pixel is a neighboring pixel of the current pixel in a prediction direction; and connect a plurality of predicted pixels to form a compensated segmentation line for the respective segmentation line. ### Claim 7 (depends on 6) The apparatus for compensating image segmentation lines according to claim 6, wherein the processor further detects a gap section in the respective segmentation line, and determines a kernel size of the classification model based on a size of the gap section. ### Claim 8 (depends on 7) The apparatus for compensating image segmentation lines according to claim 7, wherein a distance between two endpoint pixels of the detected gap section in the prediction direction of the kernel is calculated, and a length of the kernel in the prediction direction is set accordingly, so that the length is greater than the distance. ### Claim 9 (depends on 6) The apparatus for compensating image segmentation lines according to claim 6, wherein the processor further determines one or a combination of a kernel size and a padding size adopted by the classification model. ### Claim 10 (depends on 6) The apparatus for compensating image segmentation lines according to claim 6, wherein the neighboring pixel comprises a next pixel of the current pixel in the prediction direction and neighboring pixels in a vertical direction perpendicular to the prediction direction of the next pixel.