• 한국군사과학기술학회지
• /
• 제10권3호
• /
• pp.25-33
• /
• 2007

In this paper, I propose a method which can enhance the visual quality of IRST images based on a loaded location in ship and an AOS. The IRST adjusts an AOS to detect targets with various altitudes because of its narrow vertical field of view and offers various functions to enhance images with its low contrast. In the proposed method, images are divided into two regions of sea and sky on the basis of the horizon after establishing relation between an AOS and a horizon location within an image. As a result, image enhancement of the proposed method is performed adaptively according to the divided region while that of conventional method is performed for entire image without the region division. Simulation results show that the proposed method represents higher visibility compared with conventional one.

• 반도체디스플레이기술학회지
• /
• 제17권1호
• /
• pp.71-75
• /
• 2018

Image enhancement is a necessary end essential step after taking a picture with a digital camera. Many different photo software packages attempt to automate this process with various auto enhancement techniques. This paper provides and implements a system that can learn a user's preferences and apply the preferences into the process of image enhancement. Five major components are applied to the implemented system, which are computing a distance metric, finding a training set, finding an optimal parameter set, training and finally enhancing the input image. To estimate the validity of the method, we carried out user studies, and the fact that the implemented system was preferred over the method without learning user preferences.

• International Journal of Advanced Culture Technology
• /
• 제6권4호
• /
• pp.317-322
• /
• 2018

We use edge detection technique for the input image to extract the entire edges of the object in the image and then select only the edges that construct the outline of the object. By examining the positional relation between these pixels composing the outline, a simplified version of the outline of the object in the input image is generated by removing unnecessary pixels while maintaining the condition of connection of the outline. For each pixel constituting the outline, its direction is calculated by examining the positional relation with the next pixel. Then, we group the consecutive pixels with same direction into one and then change them to a line segment instead of a point. Among those line segments composing the outline of the object, a line segment whose length is smaller than a predefined minimum length of acceptable line segment is removed by merging it into one of the adjacent line segments. As a result, an outline composed of line segments of over a certain length is obtained through this process.

• Current Optics and Photonics
• /
• 제3권4호
• /
• pp.329-335
• /
• 2019

Wide FOV imaging systems are important for acquiring rich visual information. A conventional catadioptric imaging system deploys a camera in front of a curved mirror to acquire a wide FOV image. This is a cumbersome setup and causes unnecessary occlusions in the acquired image. In order to reduce both the burden of the camera deployment and the occlusions in the images, this study uses a secondary planar mirror in the catadioptric imaging system. A compact design of the catadioptric imaging system and a condition for the position of the secondary planar mirror to satisfy the central imaging are presented. The image acquisition model of the catadioptric imaging system with a secondary planar mirror is discussed based on the principles of geometric optics in this study. As a backward mapping, the acquired image is restored to a distortion-free image in the experiments.

• Journal of Information Processing Systems
• /
• 제15권1호
• /
• pp.7-21
• /
• 2019

In order to improve the edge segmentation effect of the level set image segmentation and avoid the influence of the initial contour on the level set method, a saliency level set image segmentation model based on local Renyi entropy is proposed. Firstly, the saliency map of the original image is extracted by using saliency detection algorithm. And the outline of the saliency map can be used to initialize the level set. Secondly, the local energy and edge energy of the image are obtained by using local Renyi entropy and Canny operator respectively. At the same time, new adaptive weight coefficient and boundary indication function are constructed. Finally, the local binary fitting energy model (LBF) as an external energy term is introduced. In this paper, the contrast experiments are implemented in different image database. The robustness of the proposed model for segmentation of images with intensity inhomogeneity and complicated edges is verified.

• 반도체디스플레이기술학회지
• /
• 제19권4호
• /
• pp.71-76
• /
• 2020

In this paper, we propose a super-resolution method that reconstructs compressed low-resolution images into high-resolution images. We propose a CNN model with a small number of parameters, and even if quantization is applied to the proposed model, super-resolution can be implemented without deteriorating the image quality. To further improve the quality of the compressed low-resolution image, a new degradation model was proposed instead of the existing bicubic degradation model. The proposed degradation model is used only in the training process and can be applied by changing only the parameter values to the original CNN model. In the super-resolution image applying the proposed degradation model, visual artifacts caused by image compression were effectively removed. As a result, our proposed method generates higher PSNR values at compressed images and shows better visual quality, compared to conventional CNN-based SR methods.

• 한국정보기술학회 영문논문지
• /
• 제10권1호
• /
• pp.111-120
• /
• 2020

Generative adversarial networks (GANs) achieved impressive performance on image generation and visual classification applications. However, adversarial networks meet difficulties in combining the generative model and unstable training process. To overcome the problem, we combined the deep residual network with upsampling convolutional layers to construct the generative network. Moreover, the study shows that image generation and classification performance become more prominent when the residual layers include on the generator. The proposed network empirically shows that the ability to generate images with higher visual accuracy provided certain amounts of additional complexity using proper regularization techniques. Experimental evaluation shows that the proposed method is superior to image generation and classification tasks.

• 융복합기술연구소 논문집
• /
• 제11권1호
• /
• pp.1-5
• /
• 2021

Deep learning based methods achieve state-of-the-art accuracy, however, they typically rely on supervised training with large labeled datasets. It is known in many medical applications that labeling medical images requires significant expertise and much time, and typical hand-tuned approaches for data augmentation fail to capture the complex variations in such images. This paper proposes a 3D image augmentation method to overcome these difficulties. It allows us to enrich diversity of training data samples that is essential in medical image segmentation tasks, thus reducing the data overfitting problem caused by the fact the scale of medical image dataset is typically smaller. Our numerical experiments demonstrate that the proposed approach provides significant improvements over state-of-the-art methods for 3D medical image segmentation.

• 반도체디스플레이기술학회지
• /
• 제22권1호
• /
• pp.129-133
• /
• 2023

Image compression is a fundamental technique in the field of digital image processing, which will help to decrease the storage space and to transmit the files efficiently. Recently many deep learning techniques have been proposed to promise results on image compression field. Since many image compression techniques have artifact problems, this paper has compared two deep learning approaches to verify their performance experimentally to solve the problems. One of the approaches is a deep autoencoder technique, and another is a deep convolutional neural network (CNN). For those results in the performance of peak signal-to-noise and root mean square error, this paper shows that deep autoencoder method has more advantages than deep CNN approach.

• 반도체디스플레이기술학회지
• /
• 제22권1호
• /
• pp.39-42
• /
• 2023

Recently, with the explosive increase of video streaming services, real-time live broadcasting has also increased, which leads to an infringement problem for user privacy. So, to solve such problems, we proposed the real image blurring system using dlib face-recognition library. 68 face landmarks are extracted and convert into 128 vector values. After that the proposed system tries to compare this value with the image in the database, and if it is over 0.45, it is considered as different person and image blurring processing is performed. With the proposed system, it is possible to solve the problem of user privacy infringement, and also to be utilized to detect the specific person. Through experimental results, the proposed system has an accuracy of more than 90% in terms of face recognition.