公开(公告)号：US20200167952A1

公开(公告)日：2020-05-28

申请号：US16231701

申请日：2018-12-24

Applicant: INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE

Inventor： Hsin-Yi CHEN ,  Chia-Liang YEH ,  Hsin-Cheng LIN ,  Sen-Yih CHOU

IPC: G06T7/73 ,  G06T7/90

Abstract: An object identification method includes: establishing a training data base including a photographing distance of a training image and a training camera parameter; in photographing a target test object, obtaining a test image, a depth image, an RGB image and a test camera parameter; and based on the training database, the depth image and the test camera parameter, adjusting the RGB image wherein the adjusted RGB image having a size equivalent to the training image of the training database.

公开(公告)号：US20200160545A1

公开(公告)日：2020-05-21

申请号：US16231724

申请日：2018-12-24

Applicant: INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE

Inventor： Yu-Ying LAN ,  Hsin-Cheng LIN ,  Sen-Yih CHOU

IPC: G06T7/50 ,  G06T7/73 ,  G06F3/01

Abstract: A sight vector detecting method includes: capturing a user image by an image capture unit and a depth capture unit to obtain a first image and a distance information; based on the first image and the distance information, finding an eye center location of the user; predicting a user sight location by a gaze model to find a target sight location of the user on a target; based on the eye center location of the user, finding a first word coordinate of the eye center location of the user; based on the target sight location of the user, finding a second word coordinate of the target sight location of the user; and based on the first word coordinate of the eye center location of the user and the second word coordinate of the target sight location of the user, calculating a sight vector of the user.

公开(公告)号：US20220207647A1

公开(公告)日：2022-06-30

申请号：US17135818

申请日：2020-12-28

Applicant: INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE

Inventor： Hung-Wei LIN ,  Sen-Yih CHOU ,  Hsin-Cheng LIN

IPC: G06T3/40 ,  G06T5/00 ,  G06T5/20 ,  G06T7/13 ,  G06N3/08

Abstract: A training method and training system for a resolution improvement model and a boundary detection method using the resolution improvement model are provided. The training method for the resolution improvement model includes the following steps. A low-resolution image is inputted. Pixels of the low-resolution image are captured and reorganized to generate a high-resolution image according to convolutional features. The resolution of the high-resolution image is higher than that of the low-resolution image. When capturing the low-resolution image, a condition mask is used to filter off the noise content, as well as sharpen the edge. The high-resolution image is compared with a ground-truth target image to output a discrimination result. The convolutional features are updated according to the discrimination result.

公开(公告)号：US20220147868A1

公开(公告)日：2022-05-12

申请号：US17115266

申请日：2020-12-08

Applicant: INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE

Inventor： Hsin-Cheng LIN ,  Sen-Yih CHOU

IPC: G06N20/00 ,  G06K9/62

Abstract: An adjusting method and a training system for a machine learning classification model and a user interface are provided. The machine learning classification model is used to identify several categories. The adjusting method includes the following steps. Several identification data are inputted to the machine learning classification model to obtain several confidences of the categories for each of the identification data. A classification confidence distribution for each of the identification data whose highest value of the confidences is not greater than a critical value is recorded. The classification confidence distributions of the identification data are counted. Some of the identification data are collected according to the cumulative counts of the classification confidence distributions. Whether the collected identification data belong to a new category is determined. If the collected identification data belong to a new category, the new category is added.

公开(公告)号：US20220147763A1

公开(公告)日：2022-05-12

申请号：US17184319

申请日：2021-02-24

Applicant: INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE

Inventor： Ting-Hsun CHENG ,  Yu-Ju CHAO ,  Hsin-Cheng LIN ,  Chih-Chia CHANG ,  Yu-Hsin LIN ,  Sen-Yih CHOU

IPC: G06K9/62 ,  G06T5/00 ,  G06K9/00 ,  G06T7/11 ,  G06N20/00

Abstract: A recognition system and an image augmentation and training method thereof are provided. The image augmentation and training method of a recognition system includes the following steps. A plurality of image frames are obtained, wherein each of the image frames includes an object pattern. A plurality of environmental patterns are obtained. The object pattern is separated from each of the image frames. A plurality of image parameters are set. The image frames, based on the object patterns and the environmental patterns, are augmented according to the image parameters to increase the number of the image frames. A recognition model is trained using the image frames.

公开(公告)号：US20210150272A1

公开(公告)日：2021-05-20

申请号：US16728285

申请日：2019-12-27

Applicant: INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE

Inventor： Hsin-Cheng LIN ,  Sen-Yih CHOU

IPC: G06K9/62 ,  G06N20/00 ,  G06K9/00

Abstract: An image recognition method, a training system for an object recognition model and a training method for an object recognition model are provided. The image recognition method includes the following steps. At least one original sample image of an object in a field and an object range information and an object type information in the original sample image are obtained. At least one physical parameter is adjusted to generate plural simulated sample images of the object. The object range information and the object type information of the object in each of the simulated sample images are automatically marked. A machine learning procedure is performed to train an object recognition model. An image recognition procedure is performed on an input image.