Current Issue : October - December Volume : 2019 Issue Number : 4 Articles : 5 Articles
Visual quality and algorithm efficiency are two main interests in video frame\ninterpolation. We propose a hybrid task-based convolutional neural network for fast and accurate\nframe interpolation of 4K videos. The proposed method synthesizes low-resolution frames, then\nreconstructs high-resolution frames in a coarse-to-fine fashion. We also propose edge loss, to\npreserve high-frequency information and make the synthesized frames look sharper. Experimental\nresults show that the proposed method achieves state-of-the-art performance and performs 2.69x\nfaster than the existing methods that are operable for 4K videos, while maintaining comparable\nvisual and quantitative quality....
In the current study, we were inspired by sparse analysis signal representation theory\nto propose a novel single-image super-resolution method termed â??sparse analysis-based super\nresolutionâ? (SASR). This study presents and demonstrates mapping between low and high resolution\nimages using a coupled sparse analysis operator learning method to reconstruct high resolution (HR)\nimages. We further show that the proposed method selects more informative high and low resolution\n(LR) learning patches based on image texture complexity to train high and low resolution operators\nmore efficiently. The coupled high and low resolution operators are used for high resolution image\nreconstruction at a low computational complexity cost. The experimental results for quantitative\ncriteria peak signal to noise ratio (PSNR), root mean square error (RMSE), structural similarity index\n(SSIM) and elapsed time, human observation as a qualitative measure, and computational complexity\nverify the improvements ordered by the proposed SASR algorithm....
The advancement of drones has revolutionized the production of aerial imagery. Using a\ndrone with its associated flight control and image processing applications, a high resolution\northo rectified mosaic from multiple individual aerial images can be produced within just a few\nhours. However, the positional precision and accuracy of any orthomosaic produced should not be\noverlooked. In this project, we flew a DJI Phantom drone once a month over a seven-month period\nover Oak Grove Cemetery in Nacogdoches, Texas, USA resulting in seven orthomosaics of the same\nlocation. We identified 30 ground control points (GCPs) based on permanent features in the cemetery\nand recorded the geographic coordinates of each GCP on each of the seven orthomosaics. Analyzing\nthe cluster of each GCP containing seven coincident positions depicts the positional precision of the\northomosaics. Our analysis is an attempt to answer the fundamental question, â??Are we obtaining the\nsame geographic coordinates for the same feature found on every aerial image mosaic captured by a\ndrone over time?â? The results showed that the positional precision was higher at the center of the\northomosaic compared to the edge areas. In addition, the positional precision was lower parallel to\nthe direction of the drone flight....
With the advent of 3D video compression and Internet technology, 3D videos have\nbeen deployed worldwide. Data hiding is a part of watermarking technologies and has many\ncapabilities. In this paper, we use 3D video as a cover medium for secret communication using\na reversible data hiding (RDH) technology. RDH is advantageous, because the cover image can\nbe completely recovered after extraction of the hidden data. Recently, Chung et al. introduced\nRDH for depth map using prediction-error expansion (PEE) and rhombus prediction for marking\nof 3D videos. The performance of Chung et al.â??s method is efficient, but they did not find the\nway for developing pixel resources to maximize data capacity. In this paper, we will improve the\nperformance of embedding capacity using PEE, inter-component prediction, and allowable pixel\nranges. Inter-component prediction utilizes a strong correlation between the texture image and the\ndepth map in MVD. Moreover, our proposed scheme provides an ability to control the quality of\ndepth map by a simple formula. Experimental results demonstrate that the proposed method is more\nefficient than the existing RDH methods in terms of capacity....
We present an image processing algorithm based on edge extension to correct the influence of atmospheric dispersion. The\nElden model is used to estimate the image dispersion index caused by atmospheric dispersion and the image affected by the\natmospheric dispersion is regarded as the results of the original image convolution operation. When the direct convolution is\nused to compensate the blur of star, border effect and ill-posed problem make the result unacceptable. To solve these problems,\nwe use image preprocessing and perform an edge extension method for images before the convolution. The simulated analysis\nand experimental results from a 300 mm telescope system show that the proposed method can effectively correct the influence of\natmospheric dispersion even under relatively low signal to noise ratio (SNR<2). Compared with the traditional prism correction\nand fiber correction methods, this technique can greatly reduce the complexity of the optical system....
Loading....