公开(公告)号：US10748019B2

公开(公告)日：2020-08-18

申请号：US16108430

申请日：2018-08-22

Applicant: REALTEK SEMICONDUCTOR CORP.

Inventor： Zhong-Yi Qiu ,  Wen-Tsung Huang ,  Kai Liu

IPC: G06K9/20 ,  H04N9/75 ,  G06K9/46 ,  G06K9/00 ,  G06K9/38 ,  G06T7/174 ,  G06T7/11 ,  G06T7/194

Abstract: An image processing method and an electronic apparatus for foreground image extraction include steps: (A) acquiring frame images of a dynamic image, wherein each frame image has an RGB image and an IR image; (B) acquiring an RGB image representing a dark state as an RGB capture image, acquiring an IR image representing a light state as an IR light frame image, and acquiring one IR image representing the dark state as an IR dark frame image; (C) calculating a difference image between the IR light frame image and the IR dark frame image and binarizing the difference image to generate a binarized image; and (D) acquiring a plurality of foreground pixels corresponding to an IR foreground part of the binarized image of the RGB capture image and taking the foreground pixels as the foreground image, thereby reducing the impact of ambient light, background noise and calculation.

公开(公告)号：US20180322615A1

公开(公告)日：2018-11-08

申请号：US15725291

申请日：2017-10-05

Applicant: Realtek Semiconductor Corp.

Inventor： Ching-Ju Hsiao ,  Kai Liu ,  Wen-Tsung Huang

IPC: G06T5/00 ,  G06T5/20 ,  G06T7/13

CPC classification number: G06T5/002 ,  G06T5/003 ,  G06T5/20 ,  G06T7/13 ,  G06T2207/10024 ,  G06T2207/20192

Abstract: An image processing method for processing an input image is provided. The image processing method includes the following steps: selecting a pixel of the input image; determining if the pixel is a first image edge according to at least one first calibrated pixel and at least one second pixel in the input image, in which the first calibrated pixel corresponds to at least one first pixel in the input image; and replacing a high frequency component of at least one channel of the pixel with a first calibrating high frequency component to generate a calibrated pixel if the pixel is not the first image edge, and maintaining the pixel as the calibrated pixel if the pixel is the first image edge.

公开(公告)号：US10922792B2

公开(公告)日：2021-02-16

申请号：US16732332

申请日：2020-01-01

Applicant: Realtek Semiconductor Corp.

Inventor： Zhong-Yi Qiu ,  Wen-Tsung Huang ,  Kai Liu

IPC: G06T5/00 ,  G06T5/20

Abstract: An image adjustment method includes: sequentially processing a plurality of pixels in at least one frame, wherein a pixel under processing is a current pixel, the current pixel and multiple adjacent pixels form a current block, and each current block is performed with following operations: reading a grayscale value of each pixel in the current block; determining a region grayscale value and a region variance of the current block according to the grayscale values of the pixels; generating a variance adjustment parameter via a variance adjustment function, wherein the region grayscale value is a variable of the variance adjustment function; generating an adjusted region variance according to the variance adjustment parameter and the region variance; and comparing the adjusted region variance with a variance threshold to determine whether to perform a noise suppression operation on the current pixel.

公开(公告)号：US10339635B2

公开(公告)日：2019-07-02

申请号：US15725291

申请日：2017-10-05

Applicant: Realtek Semiconductor Corp.

Inventor： Ching-Ju Hsiao ,  Kai Liu ,  Wen-Tsung Huang

IPC: G06T5/00 ,  G06T7/13 ,  G06T5/20

Abstract: An image processing method for processing an input image is provided. The image processing method includes the following steps: selecting a pixel of the input image; determining if the pixel is a first image edge according to at least one first calibrated pixel and at least one second pixel in the input image, in which the first calibrated pixel corresponds to at least one first pixel in the input image; and replacing a high frequency component of at least one channel of the pixel with a first calibrating high frequency component to generate a calibrated pixel if the pixel is not the first image edge, and maintaining the pixel as the calibrated pixel if the pixel is the first image edge.

公开(公告)号：US10528809B2

公开(公告)日：2020-01-07

申请号：US15867367

申请日：2018-01-10

Applicant: REALTEK SEMICONDUCTOR CORP.

Inventor： Kai Liu ,  Wen-Tsung Huang

IPC: G06K9/00 ,  G06F21/32 ,  A61B3/14 ,  A61B5/1171 ,  A61B5/16

Abstract: An iris image capturing device, an iris image recognition device and an iris image recognition method are provided. Multiple data sequences are captured, in which each data sequence includes an iris image. These data sequences are selected to be a positioning image or an image to be processed. The positioning image is for locating the iris, and the image to be processed is for generating a protected iris image according to where the iris is located in the positioning image.

公开(公告)号：US10943354B2

公开(公告)日：2021-03-09

申请号：US16371176

申请日：2019-04-01

Applicant: REALTEK SEMICONDUCTOR CORP.

Inventor： Zhong-Yi Qiu ,  Wen-Tsung Huang ,  Kai Liu

IPC: G06T5/00 ,  G06T7/194 ,  G06T7/11 ,  G06T7/174 ,  G06T5/50 ,  H04N5/222 ,  H04N5/235 ,  H04N5/33

Abstract: A method for extracting foreground image and electronic apparatus thereof is provided and implements infrared (IR) technology to perform the foreground image extraction to reduce ambient light effect and background noise. Furthermore, the method and electronic apparatus extract an IR light frame image indicating a light state, an IR dark frame image indicating a dark state, and a color image (e.g., RGB image, YUV image, or etc.) at different IR intensities. Then, the method and electronic apparatus calculate the relationship between the IR light frame image and the IR dark frame image to extract a better background image (including the user's face portion, body portion and hair portion) by a simple algorithm.

公开(公告)号：US10755418B2

公开(公告)日：2020-08-25

申请号：US16371203

申请日：2019-04-01

Applicant: REALTEK SEMICONDUCTOR CORP.

Inventor： Zhong-Yi Qiu ,  Wen-Tsung Huang ,  Kai Liu

IPC: H04N5/222 ,  H04N5/235 ,  H04N5/33 ,  H04N5/272 ,  H04N5/351 ,  H04N9/04 ,  G06K9/38 ,  G06K9/00 ,  G06T5/00 ,  G06T5/50 ,  G06T7/174 ,  G06T7/194 ,  G06T7/11

Abstract: A background blurred method includes steps as follows: obtaining a color image, an IR light frame image and an IR dark frame image, and calculating a differential image between the IR light frame image and the IR dark frame image to generate a binary image; obtaining a foreground image of the binary image; calculating an IR brightness mean value of the foreground image of the differential image; and filtering the color image to generate a background blurred image according to the IR brightness mean value.