公开(公告)号：US20240355106A1

公开(公告)日：2024-10-24

申请号：US18519676

申请日：2023-11-27

Applicant: Quanta Computer Inc.

Inventor： Chia-Yuan CHANG ,  Kai-Ju CHENG ,  Shao-Ang CHEN ,  Kuan-Chung CHEN

IPC: G06V10/82 ,  G06V20/70

CPC classification number: G06V10/82 ,  G06V20/70

Abstract: A method for training a segmentation model is provided. The method includes using first training images to train a segmentation model. The method includes using second training images to train an image generator. The method includes inputting real images into the segmentation model to generate predicted annotation images. The method includes inputting the predicted annotation images into the image generator to generate fake images. The method includes updating the segmentation model and the image generator according to a loss caused by differences between the real images and the fake images.

公开(公告)号：US20230298316A1

公开(公告)日：2023-09-21

申请号：US17847739

申请日：2022-06-23

Applicant: Quanta Computer Inc.

Inventor： Chia-Yuan CHANG ,  Kai-Ju CHENG ,  Yu-Hsun CHEN ,  Hao-Ping LEE ,  Tong-Ming HSU ,  Chin-Yuan TING ,  Shao-Ang CHEN ,  Kuan-Chung CHEN

IPC: G06V10/764 ,  G06V10/40 ,  G06V10/74 ,  G06V10/762

CPC classification number: G06V10/764 ,  G06V10/40 ,  G06V10/761 ,  G06V10/762

Abstract: An image classifying device is provided in the invention. The image classifying device includes a storage device, a calculation circuit and a classifying circuit. The storage device stores information corresponding to a plurality of image classes. The calculation circuit obtains a target image from an image extracting device and obtains the feature vector of the target image. The calculation circuit obtains a first estimation result corresponding to the target image based on the information corresponding to the plurality of image classes and the feature vector and obtains a second estimation result corresponding to the target image based on a reference image, wherein the reference image corresponds to one of the image classes. The classifying circuit adds the target image into one of the image classes based on the first estimation result and the second estimation result.

公开(公告)号：US20230236700A1

公开(公告)日：2023-07-27

申请号：US17819992

申请日：2022-08-16

Applicant: Quanta Computer Inc.

Inventor： Chia-Yuan CHANG ,  Jung-Wen CHANG ,  Chin-Kang CHANG ,  Ming-Yu HUANG

IPC: G06F3/042 ,  G06T7/136 ,  G06T7/73 ,  H04N5/235

CPC classification number: G06F3/0425 ,  G06T7/136 ,  G06T7/73 ,  H04N5/2354 ,  G06T2207/30241 ,  G06T2207/10152

Abstract: A method for obtaining a handwriting trajectory is provided. The method includes the following steps: capturing images of handwriting that is written with a writing brush on a piece of paper on a writing platform; obtaining positions where pixels in each of the images are lower than a threshold according to the threshold; and outputting handwriting images according to the positions.

公开(公告)号：US20210358472A1

公开(公告)日：2021-11-18

申请号：US17068044

申请日：2020-10-12

Applicant: Quanta Computer Inc.

Inventor： Chia-Yuan CHANG ,  Jung-Wen CHANG ,  Kao-Yu HSU ,  Tung-Han LEE

IPC: G10K11/34

Abstract: An inspection device is provided, including a base, a circuit assembly, an electrocardiogram sensor, a diaphragm, an annular member, and a positioning member. The base has a top surface, a bottom surface, a guiding portion, an opening, and an accommodating space. The guiding portion is formed on the top surface, the opening is formed on the bottom surface, and the accommodating space is formed between the top surface and the bottom surface. The circuit assembly is disposed in the accommodating space, and has a first contact and a second contact. The electrocardiogram sensor is disposed on the bottom surface and electrically connected to the circuit assembly. The diaphragm covers the opening. The annular member is rotatably connected to the base and has a guiding member. The guiding member is slidably connected to the guiding portion. The positioning member is affixed to the annular member and has a contacting portion.