公开(公告)号：US20220012589A1

公开(公告)日：2022-01-13

申请号：US17370585

申请日：2021-07-08

Applicant: Electronics and Telecommunications Research Institute

Inventor： Jung Jae YU ,  Jong Gook KO ,  Won Young YOO ,  Keun Dong LEE ,  Su Woong LEE ,  Seung Jae LEE ,  Yong Sik LEE ,  Da Un JUNG

IPC: G06N3/08 ,  G06N3/063 ,  G06K9/62

Abstract: A data learning device in a deep learning network characterized by a high image resolution and a thin channel at an input stage and an output stage and a low image resolution and a thick channel in an intermediate deep layer includes a feature information extraction unit configured to extract global feature information considering an association between all elements of data when generating an initial estimate in the deep layer; a direct channel-to-image conversion unit configured to generate expanded data having the same resolution as a final output from the generated initial estimate of the global feature information or intermediate outputs sequentially generated in subsequent layers; and a comparison and learning unit configured to calculate a difference between the expanded data generated by the direct channel-to-image conversion unit and a prepared ground truth value and update network parameters such that the difference is decreased.

公开(公告)号：US20190180422A1

公开(公告)日：2019-06-13

申请号：US16209577

申请日：2018-12-04

Applicant: Electronics and Telecommunications Research Institute

Inventor： Chang Woo CHU ,  Kyung Kyu KANG ,  Hye Sun KIM ,  Chang Joon PARK ,  Yun Ji BAN ,  Dong Wan RYOO ,  Jung Jae YU ,  Man Hee LEE

IPC: G06T5/00 ,  G06T7/73 ,  G06T5/50 ,  H04N5/247

Abstract: Disclosed is a method of performing geometric correction on images obtained by multiple cameras and an image obtainment apparatus therefor, the method including: receiving the images obtained by the multiple cameras; extracting feature points of the received images and checking an interrelation between the extracted feature points; estimating locations and directions of the cameras, which respectively obtain the images, by using the interrelation between the feature points; calculating a relative geometric relationship between the multiple cameras from the estimated locations and directions of the cameras when the received images are determined as the images obtained by the multiple cameras with a fixed geometric interrelationship; and performing the geometric correction on each of the images by using the calculated geometric relationship. The method of efficiently performing geometric correction according to the embodiment of the present invention enables images to be obtained easily using the unmanned device with a short photographing time.

公开(公告)号：US20140292748A1

公开(公告)日：2014-10-02

申请号：US14230834

申请日：2014-03-31

Applicant: Electronics and Telecommunications Research Institute

Inventor： Yun Ji BAN ,  Hae Dong KIM ,  Hye Sun KIM ,  Jung Jae YU ,  Kyung Ho JANG ,  Myung Ha KIM ,  Joo Hee BYON ,  Seung Woo NAM

IPC: G06T19/20 ,  G06T7/00

CPC classification number: G06T7/0075 ,  H04N13/128 ,  H04N2013/0081

Abstract: Provided is a system and method for providing a stereoscopic image by adjusting a depth value, the system including a depth value estimator to estimate a depth value of an object included in a first stereoscopic image from the stereoscopic image, a depth value adjusting unit to adjust the depth value in consideration of a display device, a stereoscopic image processing unit to process the first stereoscopic image to be a second stereoscopic image based on the adjusted depth value, and a stereoscopic image provider to provide the second stereoscopic image to the display device.

Abstract translation: 提供了一种用于通过调整深度值来提供立体图像的系统和方法，所述系统包括深度值估计器以估计来自立体图像的第一立体图像中包括的对象的深度值，深度值调整单元，用于调整 考虑到显示装置的深度值，基于调整后的深度值将第一立体图像处理成第二立体图像的立体图像处理单元，以及向显示装置提供第二立体图像的立体图像提供者。

公开(公告)号：US20140125778A1

公开(公告)日：2014-05-08

申请号：US13905824

申请日：2013-05-30

Applicant: Electronics and Telecommunications Research Institute

Inventor： Hae Dong KIM ,  Yun Ji BAN ,  Hye Sun KIM ,  Jung Jae YU ,  Kyung Ho JANG ,  Myung Ha KIM ,  Joo Hee BYON ,  Seung Woo NAM

IPC: H04N13/00

CPC classification number: H04N13/261

Abstract: Disclosed are a system for producing stereoscopic images with a hole filling algorithm, and a method thereof. A system for producing stereoscopic images according to an exemplary embodiment of the present invention includes: a movement area detecting unit examining a movement area for an object moved in an image, detecting/tracking changes of the movement area by dividing a front view and a background of the movement area, and providing result information; a filling error processing unit filling a hole detected by the movement area detecting unit and correcting a filled filling region; and a stereoscopic image visualizing unit visualizing a stereoscopic image corrected by the filling error processing unit.

Abstract translation: 公开了一种用于产生具有孔填充算法的立体图像的系统及其方法。 根据本发明的示例性实施例的用于产生立体图像的系统包括：移动区域检测单元，检查在图像中移动的物体的移动区域，通过分割前视图和背景来检测/跟踪移动区域的变化 的运动区域，并提供结果信息; 填充误差处理单元填充由所述移动区域检测单元检测到的孔，并校正填充填充区域; 以及立体图像可视化单元，可视化由填充误差处理单元校正的立体图像。