• The Journal of the Korea institute of electronic communication sciences
• /
• v.8 no.3
• /
• pp.409-414
• /
• 2013

With the increase of multimedia information such as images, researches have been realized on how to extract the high-level semantic information from low-level visual information, and a variety of techniques have been proposed to generate this information automatically. However, most of these technologies extract the semantic information between single images, it's difficult to extract semantic information when a combination of multiple objects within the image. In this paper, we extract the visual features of objects within the image and training images stored in the DB and the features of each object are defined by measuring the similarity. Using ontology reasoner, each object feature within images infers the semantic information by positional relation and associative relation. With this, it's possible to infer semantic information between objects within images, we proposed a method for inferring more complicated and a variety of high-level semantic information.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.1
• /
• pp.65-70
• /
• 2013

Recently, the field of computer vision has been actively researched through digital image which can be easily generated as the development and expansion of digital camera technology. Especially, research that extracts and utilizes the feature in image has been actively carried out. The image stitching is a method that creates the high resolution image using features extract and match. Image stitching can be widely used in military and medical purposes as well as in variety fields of real life. In this paper, we have proposed efficient image stitching method using fast feature descriptor extraction and matching based on SURF algorithm. It can be accurately, and quickly found matching point by reduction of dimension of feature descriptor. The feature descriptor is generated by classifying of unnecessary minutiae in extracted features. To reduce the computational time and efficient match feature, we have reduced dimension of the descriptor and expanded orientation window. In our results, the processing time of feature matching and image stitching are faster than previous algorithms, and also that method can make natural-looking stitched image.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2012.07a
• /
• pp.97-100
• /
• 2012

본 논문은 변환영역 기반의 방법 중 DCT 변환영역 워터마크를 이용한 이미지 무결성 검증 시스템을 제안하였다. DCT 변환영역에 워터마크를 삽입하는 것이 공간영역에 삽입하는 것보다 잡음과 일반적인 신호처리나 영상편집, 선형/비선형 필터링, 압축 등의 영상처리에 강하다. 추출된 워터마크가 원본 워터마크와의 유사도 평가에서 임계값보다 크면 이미지에 변형이 없었다고 판단한다. 만약 이미지에 변형이 있을 경우 워터마크가 훼손되어 정확한 정보를 추출하기 어려울 수 있다. 이때 추출기에서 이미지 왜곡을 역변환 시켜 워터마크를 온전하거나 최소화된 훼손으로 추출한다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.01a
• /
• pp.257-260
• /
• 2021

본 논문에서는 카메라의 포커싱과 아웃포커싱에 의해 이미지에서 뿌옇게 표현되는 DoF(Depth of field, 피사계 심도) 영역을 쿼드트리(Quadtree) 기반의 합성곱 신경망을 통해 빠르게 찾는 방법을 제안한다. 우리의 접근 방식은 RGB채널기반의 상호-상관 필터를 이용하여 DoF영역을 이미지로부터 효율적으로 분류하고, 적응형 트리인 쿼드트리를 기반으로 유의미한 영역만을 분류한다. 이 과정에서 손실 없이 온전하게 DoF영역을 추출하기 위한 필터링 과정을 거친다. 이러한 과정에서 얻어진 이미지 패치들은 전체 이미지에 비해 적은 영역으로 나타나며, 이 적은 개수의 패치들을 이용하여 네트워크 단계에서 사용할 이미지-DoF가중치 맵 데이터 쌍을 설정한다. 네트워크 과정에서 학습할 때 사용되는 데이터는 이미지와 상호-상관 필터 기반으로 추출된 DoF 가중치 맵을 이용한다. 본 논문에서 제안하는 쿼드트리 기반 합성곱 신경망은 이미지로부터 포커싱과 아웃포커싱된 DoF영역을 자동으로 추출하는 과정을 학습시키기 위해 사용된다. 결과적으로 학습에 필요한 데이터 영역이 줄어듦으로써 학습 시간과 메모리를 절약했으며, 테스트 결과로 얻은 DoF 가중치 이미지는 입력 이미지에서 DoF영역을 더욱더 빠른 시간 내에 찾아낸다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.8
• /
• pp.1941-1947
• /
• 2014

Document image binarization is largely used as previous stage of document recognition. And the result of document recognition is much affected from the result of document image binarization. There were many studies to binarize document images. The results of previous studies for document image binarization is varied according to the state of document images. In this paper, we propose a technique for document image binarization using MSER that is applied to extract objects from an image. At first, raw MSER objects are extracted from a document image. Because the raw MSER objects cannot be used for document image binarization, the extracted raw MSER objects are modified. Then the final MSER objects are used for document image binarization with the contrast image that is extracted from the document image. Experimental results show that the proposed technique is useful for document image binarization.

• Journal of Korea Game Society
• /
• v.17 no.5
• /
• pp.113-122
• /
• 2017

The location of the objects that make up a game can be an element of immersion for players. repeatedly appearing at the same position, the fun may be reduced, and as the play time elapses, the players will feel the game's fun as they appear in a larger area than at the beginning of the game play. This paper is a study to find out the location of objects according to the passage of time and to see how players controlled these objects. First, foreground images are extracted and accumulated using OpenCV programming language. The accumulated result is displayed as a heat map image. Second, the mouse movement area is detected using the mouse tracking program and compared with the heat map image, so that the screen area in which the player is interested can be known.

• The Journal of the Korea Contents Association
• /
• v.9 no.9
• /
• pp.68-79
• /
• 2009

A variance map can be used to detect and distinguish texts from background in images. However, previous variance maps work at one level and they suffer a limitation in dealing with varieties in text size, slant, orientation, translation, and color. We present a method for robustly segmenting text regions in complex color Web images using two-level variance maps. The two-level variance maps work hierarchically. The first level finds the approximate locations of text regions using global horizontal and vertical color variances with the specific mask sizes. The second level then segments each text region using intensity variance with a local mask size, which is determined adaptively. By the second process, backgrounds tend to disappear in each region and segmentation can be accurate. Highly promising experimental results have established the effectiveness of our approach.

• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.693-696
• /
• 2008

Motion blurs are caused by relative motion between the camera and the scene. The blurred image needs to be restored because undesired blur effect degrades the quality of the image. In this paper, we propose a new method for the identification of blur parameters. Experiment shows that the proposed method identifies blur extent regardless of the size of the blur and the object in the original image.

• Journal of Korea Multimedia Society
• /
• v.10 no.10
• /
• pp.1251-1259
• /
• 2007

Motion blurs are caused by relative motion between the camera and the scene. The blurred image needs to be restored because undesired blur effect degrades the quality of the image. In this paper, we propose a new method for the identification of blur extent. Experiment shows that the proposed method identifies blur extent regardless of the size of the blur and the object in the original image.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11a
• /
• pp.39-42
• /
• 2002

본 논문에서는 내용 기반 이미지 검색 및 필터링 시스템을 위한 카테고리 식별 방법을 제안한다. 제안된 방법에서는 식별 가능한 카테고리를 사전에 정의하고, 정의된 카테고리를 대표할 수 있는 이미지들을 수집한다. 다음으로, 이들로부터 다중의 내용 기반 특징값을 추출하고, 추출된 특징값들로 카테고리 데이터베이스를 구성한다. 카테고리를 식별할 질의 이미지가 입력으로 들어오면, 질의 이미지로부터 추출된 다중 특징값들을 각 카테고리의 단일 특징값과 각각 비교함으로써, 카테고리를 대표하는 다중의 유사도 거리값을 측정한다. 각 카테고리를 대표하는 다중의 유사도 거리값들은 두 가지 연산 방법에 의해 조합되는데, 조합 방법은 각각의 단일 특징값이 각 카테고리 식별에 미치는 영향을 고려하여 정의된다. 최종적으로, 각 카테고리의 조합된 유사도 거리값을 비교한 다음, 가장 유사도가 큰 카테고리를 해당 질의 이미지의 카테고리로 식별한다.