Bicubic interpolation

About: Bicubic interpolation is a research topic. Over the lifetime, 3348 publications have been published within this topic receiving 73126 citations.

Investigation on several basic interpolation methods for the use in remote sensing application

Abstract: Image from satellite is an example of remote sensing data. However, when the resolution of the available satellite image is too coarse and does not meet the required resolution, a process known as image re-sampling need to be employed, so a higher resolution version of the image could be obtained. Image re-sampling may involve interpolation, which is a process of allocating intensity value into a new generated pixel. Yet, interpolation method usually degrades the image quality. In this paper, five basic interpolation methods have been successfully implemented. These interpolation methods are nearest neighbor interpolation, bilinear interpolation, interpolation with smoothing filter, interpolation with sharpening filter, and interpolation with unsharp masking. The aim of this project is to find interpolation method that is suitable for remote sensing data. The method of our interest is the method that is easy to be implemented, but can preserve the quality of the data in term of sharpness and validness of the information. Based on the results, it is shown that all five interpolation methods tested in this research can produce good quality output when the resolution of input image is high. For low resolution input, only bilinear, smoothing filter and unsharp masking can preserve the quality of the image. However, this is only limited for interpolation with magnification factor less than 5. Bilinear, smoothing filter and unsharp masking are suitable to interpolate remote sensing data if the resolution of the input image is high enough.

Modified cubic convolution scaler with minimum loss of information

Abstract: The authors derive an adaptive version of cubic convolution interpolation for the enlargement or reduction of digital images by arbi- trary scaling factors. The adaptation is performed in each subblock (typi- cally L3L rectangular) of an image. It consists of three phases: two scaling procedures (i.e., forward and backward interpolation) and an op- timization of the interpolation kernel. In the forward interpolation phase, from the sampled data with the original resolution, we generate scaled data with different (higher or lower) resolution. The backward interpola- tion produces new discrete data by applying another interpolation to the scaled one. The phases are based on a cubic convolution interpolation whose kernel is modified to adapt to local properties of the data. During the optimization phase, we modify the parameter values to decrease the disparity between the original data and those resulting from another in- terpolation on the different-resolution output of the forward interpolating phase. The overall process is repeated iteratively. We show experimen- tal results that demonstrate the effectiveness of the proposed interpola- tion method. The algorithm exhibits significant improvement in the mini- mization of information loss when compared with the conventional interpolation algorithms. © 2001 Society of Photo-Optical Instrumentation Engineers.

Boosting Resolution and Recovering Texture of micro-CT Images with Deep Learning.

Abstract: Digital Rock Imaging is constrained by detector hardware, and a trade-off between the image field of view (FOV) and the image resolution must be made. This can be compensated for with super resolution (SR) techniques that take a wide FOV, low resolution (LR) image, and super resolve a high resolution (HR), high FOV image. The Enhanced Deep Super Resolution Generative Adversarial Network (EDSRGAN) is trained on the Deep Learning Digital Rock Super Resolution Dataset, a diverse compilation 12000 of raw and processed uCT images. The network shows comparable performance of 50% to 70% reduction in relative error over bicubic interpolation. GAN performance in recovering texture shows superior visual similarity compared to SRCNN and other methods. Difference maps indicate that the SRCNN section of the SRGAN network recovers large scale edge (grain boundaries) features while the GAN network regenerates perceptually indistinguishable high frequency texture. Network performance is generalised with augmentation, showing high adaptability to noise and blur. HR images are fed into the network, generating HR-SR images to extrapolate network performance to sub-resolution features present in the HR images themselves. Results show that under-resolution features such as dissolved minerals and thin fractures are regenerated despite the network operating outside of trained specifications. Comparison with Scanning Electron Microscope images shows details are consistent with the underlying geometry of the sample. Recovery of textures benefits the characterisation of digital rocks with a high proportion of under-resolution micro-porous features, such as carbonate and coal samples. Images that are normally constrained by the mineralogy of the rock (coal), by fast transient imaging (waterflooding), or by the energy of the source (microporosity), can be super resolved accurately for further analysis downstream.

High Accuracy Bicubic Interpolation Using Image Local Features

Abstract: In this paper, we propose a novel bicubic method for digital image interpolation. Since the conventional bicubic method does not consider image local features, the interpolated images obtained by the conventional bicubic method often have a blurring problem. In this paper, the proposed bicubic method adopts both the local asymmetry features and the local gradient features of an image in the interpolation processing. Experimental results show that the proposed method can obtain high accuracy interpolated images.