• Journal of Korea Multimedia Society
• /
• v.23 no.3
• /
• pp.412-420
• /
• 2020

This paper proposes a performance-improving face recognition system based on a super resolution method for low-resolution images. The conventional face recognition algorithm has a rapidly decreased accuracy rate due to small image resolution by a distance. To solve the previously mentioned problem, this paper generates a super resolution images based o deep learning method. The proposed method improved feature information from low-resolution images using a super resolution method and also applied face recognition using a feature extraction and an classifier. In experiments, the proposed method improves the face recognition rate when compared to conventional methods.

• Journal of the Semiconductor & Display Technology
• /
• v.19 no.4
• /
• 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.

• Journal of Broadcast Engineering
• /
• v.20 no.5
• /
• pp.706-717
• /
• 2015

The computer generated hologram (CGH) method is the technology which can generate a hologram by using only a personal computer (PC) commonly used. However, the CGH method requires a huge amount of calculational time for the 3D object with a super multi-light source or a high-definition hologram. Hence, some solutions are obviously necessary for reducing the computational complexity of a CGH algorithm or increasing the computing performance of hardware. In this paper, we propose a method which can generate a digital hologram of the 3D object with a super multi-light source using parallel distributed computing. The traditional methods has the limitation of improving CGH performance by using a single PC. However, the proposed method where a server PC efficiently uses the computing power of client PCs can quickly calculate the CGH method for 3D object with super multi-light source. In the experimental result, we verified that the proposed method can generate the digital hologram with 1,5361,536 resolution size of 3D object with 157,771 light source in 121 ms. In addition, in the proposed method, we verify that the proposed method can reduce generation time of a digital hologram in proportion to the number of client PCs.

• Journal of Broadcast Engineering
• /
• v.26 no.4
• /
• pp.429-440
• /
• 2021

With the recent development of deep convolutional neural network learning, deep learning techniques applied to single image super-resolution are showing good results. One of the existing deep learning-based super-resolution techniques is RDN(Residual Dense Network), in which the initial feature information is transmitted to the last layer using residual dense blocks, and subsequent layers are restored using input information of previous layers. However, if all hierarchical features are connected and learned and a large number of residual dense blocks are stacked, despite good performance, a large number of parameters and huge computational load are needed, so it takes a lot of time to learn a network and a slow processing speed, and it is not applicable to a mobile system. In this paper, we use the residual dense structure, which is a continuous memory structure that reuses previous information, and the residual dense channel attention block using the channel attention method that determines the importance according to the feature map of the image. We propose a method that can increase the depth to obtain a large receptive field and maintain a concise model at the same time. As a result of the experiment, the proposed network obtained PSNR as low as 0.205dB on average at 4× magnification compared to RDN, but about 1.8 times faster processing speed, about 10 times less number of parameters and about 1.74 times less computation.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.13 no.1
• /
• pp.1-8
• /
• 2018

This paper proposes a low complexity super resolution algorithm for frequency modulated continuous wave (FMCW) radar systems for foreign object debris (FOD) detection. FOD radar has a requirement to detect foreign object in small units in a large area. However, The fast Fourier transform (FFT) method, which is most widely used in FMCW radar, has a disadvantage in that it can not distinguish between adjacent targets. Super resolution algorithms have a significantly higher resolution compared with the detection algorithm based on FFT. However, in the case of the large number of samples, the computational complexity of the super resolution algorithms is drastically high and thus super resolution algorithms are difficult to apply to real time systems. In order to overcome this disadvantage of super resolution algorithm, first, the proposed algorithm coarsely obtains the frequency of the beat signal by employing FFT. Instead of using all the samples of the beat signal, the number of samples is adjusted according to the frequency of the beat signal. By doing so, the proposed algorithm significantly reduces the computational complexity of multiple signal classifier (MUSIC) algorithm. Simulation results show that the proposed method achieves accurate location even though it has considerably lower complexity than the conventional super resolution algorithms.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.11
• /
• pp.1486-1494
• /
• 2021

Character recognition is a technology required in various platforms, such as smart parking and text to speech, and many studies are being conducted to improve its performance through new attempts. However, with low-quality image used for character recognition, a difference in resolution of the training image and test image for character recognition occurs, resulting in poor accuracy. To solve this problem, this paper designed an end-to-end learning neural network that combines image super-resolution and character recognition so that the character recognition model performance is robust against various quality data, and implemented an alternative whole learning algorithm to learn the whole neural network. An alternative end-to-end learning and recognition performance test was conducted using the license plate image among various text images, and the effectiveness of the proposed algorithm was verified with the performance test.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.11a
• /
• pp.110-112
• /
• 2017

본 논문에서는 최신의 비디오 코덱 표준인 HEVC(High Efficiency Video Coding)의 화면 내 예측 부호화의 성능 향상을 위하여 SRCNN(Super Resolution Convolutional Neural Networks)을 이용하는 방법을 제안한다. SRCNN 은 비교적 최신 기술인 CNN(Convolutional Neural Network)을 사용하여 이미지를 추가적인 데이터 없이 보간 하여 해상도를 증가시키는 기술이다. HEVC 에서는 화면 내 예측의 잔차신호를 부호화 하기 위해 많은 비트를 소모하는데, 본 논문에서는 이 잔차신호들의 해상도를 낮추어 부호화 되는 비트를 줄이며, 복호화기에서 SRCNN 을 이용하여 원래의 해상도로 복원을 수행하여 압축성능을 향상 시키는 방법에 대하여 제안한다. 제안하는 기술은 HM 16.6 에 구현하였으며, CNN 트레이닝에 Caffe 라이브러리를 사용하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2011.11a
• /
• pp.106-109
• /
• 2011

본 논문에서는 연속된 프레임을 갖는 영상의 프레임간 움직임 추정 기법을 응용하여 고해상도 영상을 획득하는 초고 해상도 기법을 제안한다. 기존의 단일 영상을 이용한 초고해상도 기법의 경우 영상에서의 고주파 대역을 찾기 위해 확률 기반의 다양한 방법이 제시되었으나 연산에 사용할 수 있는 정보가 제한적이라는 문제가 존재한다. 이러한 문제를 해결하기 위해 연속된 프레임을 이용한 다양한 초고해상도 기법이 제안되었다. 본 논문에서는 주어진 영상의 전, 후의 다수 프레임을 정하여 6-tap FIR(finite impulse response) 필터를 이용하여 프레임들의 부화소(sub-pixel)를 구한 뒤에, 부화소 정밀도의 움직임 추정을 통하여 보다 정확한 고주파성분을 복원하고자 한다. 실험을 통하여 제안하는 기법이 기존의 고등차수(bi-cubic)보간법 보다 선명한 영상을 획득할 수 있었다.

• Proceedings of the KIEE Conference
• /
• 2007.10a
• /
• pp.109-110
• /
• 2007

본 논문에서는 베이시안 초해상도 영상처리시 저해상도 영상들을 고해상도 격자에 맞게 정합해서 업샘플링(upsampling)을 하는 새로운 방식에 대해 제안한다. 제안하는 업샘플링 방식은 각 장을 따로 보간하는 방식과 달리 여러 저해상도 영상의 고주파 정보가 고해상도 영상 격자의 모든 위치에 적절히 영향을 미칠 수 있도록 여러 장의 저해상도 영상의 고주파 정보를 함께 사용하여 보간한다. 보간하는 방법은 B-스플라인 (B-Spline) 기반 비정규 리샘플링(non-uniform resampling)을 기반으로 초해상도 영상처리에 맞도록 적용한다. 실험결과를 통해 일반적으로 적용되는 0-삽입(zero-padding) 업샘플링 방식과 쌍일차 보간법(bilinear interpolation) 등을 적용할 때의 효과를 살펴보고, 제안하는 방식이 일반적인 방식을 사용하는 것에 비해 정량적, 정성적으로 고해상도 정보를 더 정확히 생성해내는 것을 확인한다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.198-199
• /
• 2018

본 논문은 초해상도 기술을 위한 CNN 구조를 제안한다. 제안하는 Gradient-SR 은 고해상도 영상이 고주파 신호와 저주파 신호로 분리될 수 있다는 점을 바탕으로 고역 통과 필터인 Sobel Operator 를 CNN 기반으로 구성한다. Gradient-SR 로부터 생성된 고주파 신호는 목표 크기로 보간 된 저해상도 입력 영상과 더해짐으로 고해상도 영상을 생성한다. 실험 영상은 VDSR 이 사용 한 291 개의 영상과 B100 영상을 이용한다. 제안하는 방법은 스케일 팩터 2 에 대한 초해상도 영상 생성 실험에서 약 200%의 속도 향상을 보인다.