US 12,014,813 B2
Contouring system
Kuei-Hong Kuo, Taoyuan (TW); Yi-Ting Peng, Taoyuan (TW); Ching-Chung Kao, Taoyuan (TW); Ai-Ling Hsu, Taoyuan (TW); Yu-Ren Yang, Taoyuan (TW); Pei-Wei Shueng, Taoyuan (TW); Chun-You Chen, Taoyuan (TW); and Kuan-Chieh Huang, Taoyuan (TW)
Assigned to QUANTA COMPUTER INC., Taoyuan (TW)
Filed by Quanta Computer Inc., Taoyuan (TW)
Filed on Jun. 29, 2021, as Appl. No. 17/361,718.
Claims priority of application No. 109138746 (TW), filed on Nov. 6, 2020.
Prior Publication US 2022/0148711 A1, May 12, 2022
Int. Cl. G16H 30/20 (2018.01); G06T 7/00 (2017.01); G16H 30/40 (2018.01)
CPC G16H 30/20 (2018.01) [G06T 7/0012 (2013.01); G16H 30/40 (2018.01); G06T 2207/10081 (2013.01); G06T 2207/20084 (2013.01)] 8 Claims
OG exemplary drawing
 
1. A contouring system, comprising:
a storage unit, configured to store one or more medical images of a patient; and
a processing unit, configured to obtain a three-dimensional (3D) image stored in a multidimensional image file format from the storage unit, and execute a contouring method to generate one or more contouring images of one or more preset targets, wherein the processing unit executes the contouring method to:
use an image preprocessing module to enhance image features of the one or more preset targets by applying at least one window setting, and to improve the contouring accuracy by normalizing the scale of image intensity to an interval between 0 and 1, or by standardizing the distribution of image intensity to a mean of 0 and standard deviation of 1;
use a neural network-based contouring module to extract a plurality of image representations of the one or more preset targets from the preprocessed images, and then to generate one or more contouring images corresponding to the one or more preset targets by decoding the image representations to the contouring images with their original resolution, wherein the contouring images are binary images with an image intensity of zero or one, wherein the image intensity of zero represents a background while the image intensity of one represents the a preset target which is contoure;
a data conversion module, configured to receive one or more two-dimensional (2D) images, convert the one or more 2D images into a 3D image, and store the 3D image as a file in the multidimensional image file format, wherein a file format of the one or more 2D image includes a common image file format or a multidimensional image file format, wherein the multidimensional image file format follows Digital Imaging and Communication in Medicine (DICOM) standard, wherein DICOM images include a metadata and a pixel data; and
a data de-identification module, configured to retain an image resolution, an affine transformation matrix, and a numerical scale transformation of an image header in the multidimensional image file format and thereby remove the patient's personal data from the image header in the multidimensional image file format using a data de-identification method, and store both an image data and the de-identified image header as another image in the multidimensional image file format.