Granted Patent B2

US 12,322,194 · App. 18/088,988 · Granted Jun 3, 2025

Inventors: Tzu-Kuei Shen (Hsinchu, TW); Linda Siana (Hsinchu, TW); Guang-Hao Suen (Hsinchu, TW); Liang-Wei Sheu (Hsinchu, TW); Chien-Ting Yang (Hsinchu, TW); Yuh-Min Chen (Taipei, TW); Heng-sheng Chao (Taipei, TW); Chung-Wei Chou (Taipei, TW); Tsu-Hui Shiao (Taipei, TW); Yi-Han Hsiao (Taipei, TW); Chi-Lu Chiang (Taipei, TW)

Assignees: V5MED INC.; TAIPEI VETERANS GENERAL HOSPITAL

G06V20/693G06T7/0012G06V20/698G06T2207/10056G06T2207/30024

Abstract

Claims (19)

1. A method for determining characteristic cells based on image recognition, scanning a stained tissue solution on a microslide to recognize the characteristic cells of the stained tissue solution, and the method comprising:

by a scanning device, capturing a full image of the microslide, wherein the scanning device comprises a microscope and camera;

by a host, receiving the full image and selecting a plurality of images that comprise stained blocks from the full image with an image processing algorithm;

by the host, sequentially performing image recognition on the plurality of images to recognize a plurality of stained cells of the plurality of images, and determining whether the plurality of stained cells comprise the characteristic cells with an artificial intelligence (AI) model; and

selecting a plurality of interested images from the plurality of images that comprise the characteristic cells, transforming a coordinate system of the plurality of interested images into an original coordinate system of the full image, and employing the scanning device to capture the plurality of interested images along a Z axis of the original coordinate system of the full image, thereby obtaining and outputting a plurality of sets of pictures,

wherein when the AI model determines that the plurality of stained cells comprise the characteristic cells, the characteristic cells are scored and the plurality of interested images are selected according to scores of the characteristic cells, and

wherein the characteristic cells are scored according to sizes and integrity of the characteristic cells.

2. The method for determining the characteristic cells based on image recognition according to claim 1 , wherein the scanning device further comprises an electric stage, the microscope is mounted on the electric stage, the camera is mounted on the microscope, and microslide is mounted on the electric stage.

3. The method for determining the characteristic cells based on image recognition according to claim 2 , wherein the microscope scans the microslide and then employs the camera to obtain the full image of the microslide after adjusting a magnification.

4. The method for determining the characteristic cells based on image recognition according to claim 1 , wherein the step of determining whether the plurality of stained cells comprise the characteristic cells with the AI model comprises:

the AI model determines whether the characteristic cells comprised by the plurality of stained cells are abnormal when the plurality of stained cells comprise the characteristic cells; and

the AI model determines whether the plurality of stained cells in a next image comprise the characteristic cells when the plurality of stained cells do not comprise the characteristic cells.

5. A system for determining characteristic cells based on image recognition, configured to scan a stained tissue solution on a microslide to recognize the characteristic cells of the stained tissue solution, comprising:

a scanning device comprising an electric stage, a microscope, and a camera, wherein the microscope is mounted on the electric stage and configured to display a full image of the microslide, and the camera is mounted on a lens of the microscope and configured to capture the full image; and

a host electrically connected to the camera, wherein the host is configured to receive the full image, select a plurality of images that comprise stained blocks from the full image with an image processing algorithm, sequentially perform image recognition on the plurality of images to recognize a plurality of stained cells of the plurality of images, determine whether the plurality of stained cells comprise the characteristic cells with an artificial intelligence (AI) model, select a plurality of interested images from the plurality of images that comprise the characteristic cells, transform a coordinate system of the plurality of interested images into an original coordinate system of the full image, and employ the scanning device to capture the plurality of interested images along a Z axis of the original coordinate system of the full image, thereby obtaining and outputting a plurality of sets of pictures,

wherein when the AI model determines that the plurality of stained cells comprise the characteristic cells, the characteristic cells are scored and the plurality of interested images are selected according to scores of the characteristic cells, and

wherein the characteristic cells are scored according to sizes and integrity of the characteristic cells.

6. The system for determining the characteristic cells based on image recognition according to claim 5 , wherein the microscope scans the microslide and then employs the camera to obtain the full image of the microslide after adjusting a magnification.

7. The system for determining the characteristic cells based on image recognition according to claim 5 , wherein the AI model determines whether the characteristic cells comprised by the plurality of stained cells are abnormal when the plurality of stained cells comprise the characteristic cells, and the AI model determines whether the plurality of stained cells in a next image comprise the characteristic cells when the plurality of stained cells do not comprise the characteristic cells.

Assignments (2)

Priority Claims (1)

Continuity (1)

References Cited (25)