• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.8
• /
• pp.1121-1129
• /
• 2022

Object recognition is carried out using RGB images in various object recognition studies. However, RGB images in dark illumination environments or environments where target objects are occluded other objects cause poor object recognition performance. On the other hand, IR images provide strong object recognition performance in these environments because it detects infrared waves rather than visible illumination. In this paper, we propose an RGB-IR fusion model, high-frequency interchange network (HINet), which improves object recognition performance by combining only the strengths of RGB-IR image pairs. HINet connected two object detection models using a mutual high-frequency transfer (MHT) to interchange advantages between RGB-IR images. MHT converts each pair of RGB-IR images into a discrete cosine transform (DCT) spectrum domain to extract high-frequency information. The extracted high-frequency information is transmitted to each other's networks and utilized to improve object recognition performance. Experimental results show the superiority of the proposed network and present performance improvement of the multispectral object recognition task.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.301-307
• /
• 2024

This study introduces a novel cascade fusion architecture aimed at enhancing thermal images across various scale conditions. The processing of thermal images at multiple scales has been challenging due to the limitations of existing methods that are designed for specific scales. To overcome these limitations, this paper proposes a unified framework that utilizes cascade feature fusion to effectively learn multi-scale representations. Confidence maps from different image scales are fused in a cascaded manner, enabling scale-invariant learning. The architecture comprises end-to-end trained convolutional neural networks to enhance image quality by reinforcing mutual scale dependencies. Experimental results indicate that the proposed technique outperforms existing methods in multi-scale thermal image enhancement. Performance evaluation results are provided, demonstrating consistent improvements in image quality metrics. The cascade fusion design facilitates robust generalization across scales and efficient learning of cross-scale representations.

• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.535-540
• /
• 2006

집적 영상(Integral Imaging) 시스템은 관찰자가 특수안경의 착용 없이 육안으로 3 차원 영상을 볼 수 있는 무안경식 양안시차 디스플레이 방식 중 하나로, 수직, 수평 시차와 총천연색의 영상을 제공한다. 집적영상 시스템은 3 차원 정보를 2 차원 엘리멘탈 이미지 (Elemental image)의 형태로 저장하는데, 엘리멘탈 이미지는 조금씩 다른 방향에서 제한된 크기로 촬영된 이미지이다. 엘리멘탈 이미지는 컴퓨터 그래픽으로 만들어질 수도 있는데, 이를 이용하는 집적 영상 방식을 CG 직접 영상 시스템이라 한다. 이와 같이 컴퓨터 계산에 의해 엘리멘탈 이미지를 얻는 과정을 이미지 매핑 (Image mapping)이라 부른다. 이제까지 제안된 이미지 매핑 방식에는 점대점 (Point to Point), MVR (Multi-Viewpoint Rendering), PGR (Parallel Group Rendering) 이 있다. 그러나 이런 방식들은 계산량이 많거나 렌즈 어레이 개수의 증가에 의해 속도에 영향을 받는 단점이 있어, 아직 가상현실 같은 실시간 CG 응용 분야에 사용하기 어려운 문제가 있다. 본 논문에서는 VVR (Viewpoint Vector Rendering)이라는 기존의 방법과 비교해 향상된 새로운 이미지 매핑 방법을 제안한다. 먼저 VVR 개념을 자세히 설명한 후 VVR 을 사용한 집적 영상 시스템을 구현하여 MVR 방법과 비교 분석한 실험결과와 개선되어야 할 방향을 제시한다.

• The KIPS Transactions:PartB
• /
• v.13B no.3 s.106
• /
• pp.231-238
• /
• 2006

In this paper, the method to make lossless image compression that holds considerable part of total volume of mobile game has been proposed. To increase the compression rate, we compress the image by Deflate algorithm defined in RFC 1951 after reorganize it at preprocessing stage before conducting actual compression. At the stage of preprocessing, we obtained the size of a dictionary based on the information of image which is the feature of Dictionary-Based Coding, and increased the better compression rate than compressing in a general manner using in a way of restructuring image by pixel packing method and DPCM prediction technique. It has shown that the method increased 9.7% of compression rate compare with existing mobile image format, after conducting the test of compression rate applying the suggested compression method into various mobile games.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.701-704
• /
• 2020

이미지 초해상도는 딥러닝의 발전과 함께 이를 활용하며 눈에 띄는 성능향상을 이루었다. 딥러닝을 기반으로 한 대부분의 이미지 초해상도 연구는 딥러닝 네트워크 모델의 구조에 대한 연구 위주로 진행되어 왔다. 그러나 최근 들어 딥러닝 기반의 이미지 초해상도가 합성된 데이터에 대해서는 높은 성능을 보이지만 실제 데이터에 대해서는 높은 성능을 보이지 못한다는 사실이 주목받고 있다. 이에 따라 모델 구조를 바꿔 성능을 향상 시키는 것에는 한계가 있어 데이터의 활용이나 학습 방법에 대한 연구의 필요성이 증대되고 있다. 따라서 본 논문은 이미지 초해상도를 위한 난이도 조절 기반 전이학습법(transfer learning)을 제안한다. 제안된 방법에서는 이미지 초해상도를 배율을 난이도가 쉬운 낮은 배율부터 순차적으로 전이학습을 진행한다. 이는 이미지 초해상도의 배율이 높아질수록 학습이 어렵기 때문이다. 결과적으로 본 논문에서는 높은 배율의 이미지 초해상도를 진행하기 위해 낮은 배율의 이미지 초해상도, 즉 난이도가 쉬운 학습부터 점진적으로 학습을 진행하였을 때 더욱 빠르고 효과적으로 학습할 수 있음을 보여준다. 제안된 전이학습 방법을 통해 적은 횟수의 업데이트로 학습을 진행하였을 때 일반적인 학습방법 대비 약 0.18 dB 의 PSNR 상승을 얻어, RealSR [9] 데이터셋에서 28.56 dB의 성능으로 파라미터 수 대비 높은 성능을 얻을 수 있었다.

• Journal of Convergence for Information Technology
• /
• v.12 no.1
• /
• pp.1-6
• /
• 2022

As augmented reality technology becomes more common and prevelant, marker-based AR contents are applied in various ways. However AR contents are still hardly utilized due to the low recognition rate of marker images. In order to increase the recognition rate of AR marker images, this paper experiment and analyzed how much the recognition rate of markers could be improved when image correction and design changes was applied. The experimental result shows that the image correction task could significantly improve the number of image characteristics and the recognition grade if the image was modified in a way its saturation value is increased. Moreover, the recognition rate was improved even more when regular pattern design was added to the original marker image. In conclusion, it was possible to make the marker well recognized through proper correction of the image and additional process of pattern design in the process of producing the marker image.

• Proceedings of the Korean Information Science Society Conference
• /
• 2008.06c
• /
• pp.348-353
• /
• 2008

이미지의 사용이 증가함에 따라 이미지 중 사용자가 원하는 이미지를 효율적으로 검색하기 위한 방법들이 연구되어 왔다. 본 논문에서는 질의 이미지를 분석하여 이미지 특징(feature)을 추출한 후 이미지 특징에 대한 유사도 평가를 통한 이미지 검색 및 온톨로지를 기반으로 검색된 이미지들과 유사하다고 판단된 이미지와 그러한 이미지들의 의미적 정보를 추출하는 방법을 제안한다. 제안된 시스템은 질의 이미지에서 색상, 질감, 모양 등의 특징을 추출하여 유사도 평가를 통해 검색된 이미지를 제공하고, 내용기반 이미지 검색 방식을 통해 이미지를 검색하고, 온톨로지를 이용해 이미지의 의미적 정보를 추출하여 사용자에게 이미지와 관련된 의미적 정보를 제공한다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2003.05a
• /
• pp.171-174
• /
• 2003

This paper is a study on the 3D infrared image enhancement with Stereoscopic algorithm on still infrared image. The adapted stereo method is that the depth is extracted by comparison with right-left image, and the enhanced 3D infrared image by matching based on feature is realized. As the result of experiment this method forced the more smooth edge lines of 3D infrared images.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.3
• /
• pp.577-581
• /
• 2003

The 3D infrared image enhancement with Stereoscopic algorithm on still infrared image is presented. The adapted stereoscopic method is that the depth is extracted by comparison with right-left image, and the enhanced 3D infrared image by matching based on feature is realized. As the result of experiment this method forces a 3D infrared image by the edge lines to be more smooth.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.12 no.5
• /
• pp.163-175
• /
• 2017

The purpose of this study is to find out how to improve the brand image and loyalty of cafe by recognizing that social culture of middle school cafe, which is an important service quality in cafe establishment, emerges as a main characteristic of new coffee business. First, reliability, confidentness, professionalism, accessibility, and socio - culturality of the service quality of coffee specialty shops improve brand image. Confidence and professionalism play an important role in enhancing brand loyalty, and brand image has a significant effect on brand loyalty. Respectively, Among the service quality, social culture has a strong influence on brand image but it is not a direct influence on brand loyalty. Second, in the relationship between brand loyalty of coffee service quality, brand image shows full mediation effect on reliability, partial mediation effect on confidence, professionalism, accessibility, socialcultural property, and mediation effect on response and empathy. Third, as a result of analyzing the moderating effects of coffee shop types on the relationship between service quality and brand image of coffee specialty shops, reliability, confidentiality, and accessibility are positive factors in the nationwide franchise. On the other hand, in the private $caf{\acute{e}}$, professionalism and socio-culturality are the main factors for improving the brand image. In the case of the local franchise, similar to the franchise in the country, the improvement of service quality for responsiveness and professionalism is a positive factor Respectively. As a result, nationwide franchise $caf{\acute{e}}s$ have priority in enhancing brand image and brand loyalty through accessibility and assurance of service quality. On the other hand, in case of local franchise $caf{\acute{e}}$, it can be seen that the service quality is enhanced and the brand image and brand loyalty can be further improved through service professionalism and accessibility. On the other hand, regional cafes are more important than national franchises or local franchise cafes, and a strategy to enhance customer loyalty is needed through service strategies emphasizing socio - cultural aspects.