• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.540-541
• /
• 2018

In this paper, we propose an super-resolution method that improves resolution by using DWT (Discrete Wavelet Transform) for low resolution infra-red images. In this method, DWT is performed in a manner that does not reduce the resolution of an image input through an infra-red camera to generate sub-bands of the same resolution (LH, HL, and HH) And the original infra-red image is used to perform an inverse-DWT to obtain an infra-red image with improved resolution. Experimental results show that the mean SSIM value of the proposed method is 0.989861, which is about 0.004 higher than that of the conventional Bi-linear and Bi-cubic filters.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.7
• /
• pp.2328-2344
• /
• 2022

2D human pose estimation still faces difficulty in low-resolution images. Most existing top-down approaches scale up the target human bonding box images to the large size and insert the scaled image into the network. Due to up-sampling, artifacts occur in the low-resolution target images, and the degraded images adversely affect the accurate estimation of the joint positions. To address this issue, we propose a multi-resolution input feature fusion network for human pose estimation. Specifically, the bounding box image of the target human is rescaled to multiple input images of various sizes, and the features extracted from the multiple images are fused in the network. Moreover, we introduce a guiding channel which induces the multi-resolution input features to alternatively affect the network according to the resolution of the target image. We conduct experiments on MS COCO dataset which is a representative dataset for 2D human pose estimation, where our method achieves superior performance compared to the strong baseline HRNet and the previous state-of-the-art methods.

• Journal of Information Processing Systems
• /
• v.18 no.1
• /
• pp.12-25
• /
• 2022

In recent years, deep convolutional neural networks have made significant progress in the research of single image super-resolution. However, it is difficult to be applied in practical computing terminals or embedded devices due to a large number of parameters and computational effort. To balance these problems, we propose CSRNet, a lightweight neural network based on channel split residual learning structure, to reconstruct highresolution images from low-resolution images. Lightweight refers to designing a neural network with fewer parameters and a simplified structure for lower memory consumption and faster inference speed. At the same time, it is ensured that the performance of recovering high-resolution images is not degraded. In CSRNet, we reduce the parameters and computation by channel split residual learning. Simultaneously, we propose a double-upsampling network structure to improve the performance of the lightweight super-resolution network and make it easy to train. Finally, we propose a new evaluation metric for the lightweight approaches named 100_FPS. Experiments show that our proposed CSRNet not only speeds up the inference of the neural network and reduces memory consumption, but also performs well on single image super-resolution.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.9
• /
• pp.3745-3761
• /
• 2020

Low-rank representation methods already achieve many applications in the image reconstruction. However, for high-gradient image patches with rich texture details and strong edge information, it is difficult to find sufficient similar patches. Existing low-rank representation methods usually destroy image critical details and fail to preserve edge structure. In order to promote the performance, a new representation-based image super-resolution reconstruction method is proposed, which combines gradient domain guided image filter with the structure-constrained low-rank representation so as to enhance image details as well as reveal the intrinsic structure of an input image. Firstly, we extract the gradient domain guided filter of each atom in high resolution dictionary in order to acquire high-frequency prior information. Secondly, this prior information is taken as a structure constraint and introduced into the low-rank representation framework to develop a new model so as to maintain the edges of reconstructed image. Thirdly, the approximate optimal solution of the model is solved through alternating direction method of multipliers. After that, experiments are performed and results show that the proposed algorithm has higher performances than conventional state-of-the-art algorithms in both quantitative and qualitative aspects.

• Journal of Broadcast Engineering
• /
• v.18 no.3
• /
• pp.463-473
• /
• 2013

In this paper, we propose a super-resolution (SR) image reconstruction algorithm using multi-view images. We acquire 25 images from multi-view cameras, which consist of a $5{\times}5$ array of cameras, and then reconstruct an SR image of the center image using a low resolution (LR) input image and the other 24 LR reference images. First, we estimate disparity maps from the input image to the 24 reference images, respectively. Then, we interpolate a SR image by employing the LR image and matching points in the reference images. Finally, we refine the SR image using an iterative regularization scheme. Experimental results demonstrate that the proposed algorithm provides higher quality SR images than conventional algorithms.

• Spatial Information Research
• /
• v.13 no.4 s.35
• /
• pp.345-354
• /
• 2005

Super resolution image reconstruction method refers to image processing algorithms that produce a high resolution(HR) image from observed several low resolution(LR) images of the same scene. This method has proven to be useful in many practical cases where multiple frames of the same scene can be obtained, such as satellite imaging, video surveillance, video enhancement and restoration, digital mosaicking, and medical imaging. In this paper, we applied the super resolution reconstruction method in spatial domain to video sequences. Test images are adjacently sampled images from continuous video sequences and are overlapped at high rate. We constructed the observation model between the HR images and LR images applied with the Maximum A Posteriori(MAP) reconstruction method which is one of the major methods in the super resolution grid construction. Based on the MAP method, we reconstructed high resolution images from low resolution images and compared the results with those from other known interpolation methods.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.46 no.5
• /
• pp.39-48
• /
• 2009

Many super-resolution reconstruction algorithms have been proposed since it was the first proposed in 1984. The spatial domain approach of the super-resolution reconstruction methods is accomplished by mapping the low resolution image pixels into the high resolution image pixels. Generally, a super-resolution reconstruction algorithm by using the spatial domain approach has the noise problem because the low resolution images have different noise component, different PSF, and distortion, etc. In this paper, we proposed the new super-resolution reconstruction method that uses the L1 norm to minimize noise source and also uses the Huber norm to preserve edges of image. The proposed algorithm obtained the higher image quality of the result high resolution image comparing with other algorithms by experiment.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.12
• /
• pp.1604-1611
• /
• 2020

Light field image captured by a microlens array-based camera has many limitations in practical use due to its low spatial resolution and angular resolution. High spatial resolution images can be easily acquired with a single image super-resolution technique that has been studied a lot recently. But there is a problem in that high angular resolution images are distorted in the process of using disparity information inherent among images, and thus it is difficult to obtain a high-quality angular resolution image. In this paper, we propose light field angular super-resolution that extracts an initial feature map using an dilated convolutional neural network in order to effectively extract the view difference information inherent among images and generates target image using a residual neural network. The proposed network showed superior performance in PSNR and subjective image quality compared to existing angular super-resolution networks.

• Korean Journal of Remote Sensing
• /
• v.37 no.6_1
• /
• pp.1573-1587
• /
• 2021

A Drone image is an ultra-high-resolution image that is several or tens of times higher in spatial resolution than a satellite or aerial image. Therefore, drone image-based remote sensing is different from traditional remote sensing in terms of the level of object to be extracted from the image and the amount of data to be processed. In addition, the optimal scale and size of data used for model training is different depending on the characteristics of the applied deep learning model. However, moststudies do not consider the size of the object to be found in the image, the spatial resolution of the image that reflects the scale, and in many cases, the data specification used in the model is applied as it is before. In this study, the effect ofspatial resolution and image size of drone image on the accuracy and training time of the semantic segmentation deep learning model of six wintering vegetables was quantitatively analyzed through experiments. As a result of the experiment, it was found that the average accuracy of dividing six wintering vegetablesincreases asthe spatial resolution increases, but the increase rate and convergence section are different for each crop, and there is a big difference in accuracy and time depending on the size of the image at the same resolution. In particular, it wasfound that the optimal resolution and image size were different from each crop. The research results can be utilized as data for getting the efficiency of drone images acquisition and production of training data when developing a winter vegetable segmentation model using drone images.

• Journal of Korean Society for Geospatial Information Science
• /
• v.10 no.5 s.23
• /
• pp.45-51
• /
• 2002

The goal of image merging techniques are to enhance the resolution of low-resolution images using the detail information of the high-resolution images. Among the several image merging methods, wavelet-based image merging techniques have the advantages of efficient decorrelation of image bands and time-scale analysis. However, they have no regard for spatial information between the bands. In other words, multiresolution data merging methods merge the same information-the detail information of panchromatic image-with other band images, without considering specific characteristics. Therefore, a merged image contains much unnecessary information. In this paper, we discussed this 'mixing' effect and, proposed a method to classify the detail information of the panchromatic image according to the spatial and spectral characteristics, and to minimize distortion of the merged image.