Electronics and Telecommunications Research Institute

National University of Sciences & Technology

Ideahub Inc.

Information Technology University (itu)

Yong Ju Cho

Ji Hun Cha

Joo Myoung Seok

Seong Yong Lim

Sang Woo Ahn

Jeong Il Seo

Rehan Hafiz

Jin Woong Kim

Yong Ju Lee

Seung Kwon Beack

Soon Heung Jung

Myung Seok Ki

Arshad Ali

Jin Woo Hong

Ashar Rasul

Muhammad Murtaza Khan

Jae Gon Kim

Jeongil Seo

Hyun Cheol Kim

Young Kwon Lim

Hee Kyung Lee

Sangwoo Ahn

Seung Jun Yang

Hyoung Joong Kim

Soon-heung Jung

Han Kyu Lee

Ji Sung Yoo

Sung Soo Hwang

Mohsen Ali

Muhammad Faisal

Seong Dae Kim

Muhammad Murtaza

Moon Sup Song

Saleh Saeed

Aman Irshad

Muhammad Umer Kakli

Unsang Park

Atif Ahmed

Wook Joong Kim

Haris Anis

Injae Lee

Mutahir Latif

Disclosed herein is an image deep learning model training method. The method includes sampling a twin negative comprising a first negative sample and a second negative sample by selecting the first negative sample with a highest similarity out of an anchor sample and a positive sample constituting a matching pair in each class and by selecting the second negative sample with a highest similarity to the first negative sample, and training the samples to minimize a loss of a loss function in each class by utilizing the anchor sample, the positive sample, the first and second negative samples for each class. The first negative sample is selected in a different class from a class comprising the matching pair, and the second negative sample is selected in a different class from classes comprising the matching pair and the first negative sample.

Method of training image deep learning model and device thereof

The present invention provides a method for image preprocessing, the method including: identifying a plurality of input images to process image registration; selecting at least one transformation method among histogram equalization and gamma transformation in consideration of hue and brightness values of the plurality of input images; processing the histogram equalization on the plurality of input images in response to the histogram equalization method being selected; and processing the gamma transformation on the plurality of input images, in response to the gamma transformation method being selected.

Method and apparatus for image preprocessing

The present invention relates to an apparatus and a method of stitching a real-time image. More specifically, stitching-related information is generated, a parallax occurrence region in an overlap region is tracked, and correction for reducing parallax occurring in the parallax occurrence region and correction of an image distortion occurring in a non-overlap region are performed for output.

Apparatus and method of parallax-minimized stitching by using HLBP descriptor information

Disclosed is an apparatus and method of generating a stitched image. A method of generating a stitched image according to the present disclosure includes: obtaining a plurality of input images; determining whether or not a parallax problem occurs between overlapping areas of the obtained input images; updating a predetermined look-up table according to the determination; and generating a stitched image by applying the updated look-up table to the input images.

Apparatus and method of generating stitched image based on look-up table

A method of receiving content in a client is provided. The method may include receiving, from a server, a spatial set identifier (ID) corresponding to a tile group including at least one tile, sending, to the server, a request for first content corresponding to metadata, and receiving, from the server, the first content corresponding to the request.

Apparatus and method for providing content

Disclosed is an apparatus and method for processing information of multiple cameras. The apparatus for processing information of multiple cameras according to an embodiment of the present disclosure includes: multiple cameras obtaining an image containing at least one object; a location information identification unit identifying location information between the at least one object and one of the multiple cameras, the location information identification unit being provided in each of the multiple cameras, wherein the location information comprises a device identifier identifying the location information identification unit and a location data between the at least one object and one of the multiple cameras; and an image processor processing image information obtained from the multiple cameras by using the location information received from the location information identification unit and converting the device identifier received from the location information identification unit into an object identifier identifying the at least one object.

Apparatus and method for processing information of multiple cameras

Rate adaptation is carried out using bit error rate (BER) to enable effective multimedia transmission. The BER can be estimated using signal strength in a MAC layer and modulation information (FIGS.-), and can be compatibly used in different wireless networks by means of message standardization.

Rate adaptation method using bit error rate for multimedia service and apparatus therefor

Provided is a parallax minimization stitching method and apparatus using control points in an overlapping region. A parallax minimization stitching method may include defining a plurality of control points in an overlapping region of a first image and a second image received from a plurality of cameras, performing a first geometric correction by applying a homography to the control points, defining a plurality of patches based on the control points, and performing a second geometric correction by mapping the patches.

Parallax minimization stitching method and apparatus using control points in overlapping region

Disclosed are a method and an apparatus for generating a panoramic image based on an image quality. The method of generating a panoramic image includes extracting a matching point to connect a base image captured by a first camera and a reference image captured by a second camera, geometrically transforming the reference image by determining a homography between the base image and the reference image, determining a change in image quality of the geometrically transformed reference image, and generating a panoramic image in which the geometrically transformed reference image is connected to the base image based on the determined change in image quality.

Apparatus and method for generating panoramic image based on image quality

An apparatus and a method for displaying a panoramic image using a look-up table (LUT) are disclosed, including generating an LUT may include determining first geometric correction information to transform an input domain pixel coordinate system of an input image to a panorama domain pixel coordinate system of a panoramic image, determining second geometric correction information to transform an output domain pixel coordinate system of an output image of the panoramic image to the panorama domain pixel coordinate system of the panoramic image, determining third geometric correction information to transform the output domain pixel coordinate system of the output image to the input domain pixel coordinate system of the input image based on the first geometric correction information and second geometric correction information, and generating an LUT that maps the output domain pixel coordinate system of the output image to the input domain pixel coordinate system of the input image.

System and method for displaying panoramic image using single look-up table

Provided is an apparatus and method for detecting a key point using a high-order Laplacian of Gaussian (LoG) kernel. The high-order LoG kernel is generated based on an LoG operator which is calculated by sequentially differentiating an LoG operator with respect to x and y of an image. A scale space is generated based on the high-order LoG kernel and the key point is detected by comparing a current pixel in the scale space to pixels adjacent to the current pixel.

Apparatus and method for detecting key point using high-order laplacian of gaussian (LoG) kernel

The present specification is related to a method for performing multi-projection and a multi-projection system for minimizing a black offset in a multi-projection environment. The present specification provides the method for performing multi-projection comprising estimating the intensity transfer function (ITF) of a plurality of projectors, calculating an optimal black offset threshold for each of the projectors by using each of the ITFs, and applying the optimal black offset threshold to an image projection by each of the projectors.

Method and system for multi-projection

Growing community of inventors

Daejeon, South Korea

Yong Ju Cho