Pixel Replication Research Articles

Combining morphological filtering, anisotropic diffusion and block-based data replication for automatically detecting and recovering unscanned gaps in remote sensing images

Filling damaged pixels in satellite images is a key task present in many Remote Sensing applications. As a representative example of image restoration issue, we can refer to the failure of the Scan Line Corrector (SLC) on board the Landsat Enhanced Thematic Mapper Plus (ETM +) sensor, in which 22% of the scanned pixels in the SLC-off images were missed, thus creating unexpected stipe-type gaps in the scenes. In order to improve the usability of ETM + SLC-off data in a straightforward manner, in this paper we propose a unified methodology that automatically segments and repairs Landsat-7 scenes occluded by stripes. The proposed framework combines Morphology-based filtering, anisotropic diffusion and block-based pixel replication as an effective, fully unsupervised restoration methodology designed to cope with different gap sizes in Landsat images. Our approach does not require having as input data any prior gap mask, side reference image or time-dependent frames of the same scene to work properly. As shown in the experimental results, the current methodology performs adequately for a variety of multispectral remote sensing images with different stripe-size thicknesses and heterogeneous segments. We attest to the accuracy and robustness of our end-to-end framework throughout a variety of qualitative and quantitative evaluations involving state-of-the-art restoration methods.

Separating Touching Cells Using Pixel Replicated Elliptical Shape Models.

One of the most important and error-prone tasks in biological image analysis is the segmentation of touching or overlapping cells. Particularly for optical microscopy, including transmitted light and confocal fluorescence microscopy, there is often no consistent discriminative information to separate cells that touch or overlap. It is desired to partition touching foreground pixels into cells using the binary threshold image information only, and optionally incorporating gradient information. The most common approaches for segmenting touching and overlapping cells in these scenarios are based on the watershed transform. We describe a new approach called pixel replication for the task of segmenting elliptical objects that touch or overlap. Pixel replication uses the image Euclidean distance transform in combination with Gaussian mixture models to better exploit practically effective optimization for delineating objects with elliptical decision boundaries. Pixel replication improves significantly on commonly used methods based on watershed transforms, or based on fitting Gaussian mixtures directly to the thresholded image data. Pixel replication works equivalently on both 2-D and 3-D image data, and naturally combines information from multi-channel images. The accuracy of the proposed technique is measured using both the segmentation accuracy on simulated ellipse data and the tracking accuracy on validated stem cell tracking results extracted from hundreds of live-cell microscopy image sequences. Pixel replication is shown to be significantly more accurate compared with other approaches. Variance relationships are derived, allowing a more practically effective Gaussian mixture model to extract cell boundaries for data generated from the threshold image using the uniform elliptical distribution and from the distance transform image using the triangular elliptical distribution.

Effect of Non-Integer Scaling on the Recovery of Data-bearing Marks

The spatial quantization imparted by printed pixels becomes significant when printing finely detailed bitonal images such as data-bearing halftones. This paper explores the consequences non-integer scaling has on data recovery error from such data-bearing marks. We verify that different printer manufacturers use nearest-neighbor scaling, and conducted hundreds of print and mobile camera capture measurements to quantify data recovery errors as a function of printed pixel replication factor. We tested the effects of multiple payloads represented by several data-bearing images, printed on a variety of printers, and captured at different camera distances. The analog print–and-capture experiments are compared with digital simulations, using several error measures. The results support a surprising conclusion: there is no significant advantage in forcing printed pixel replication to be an integer.

33.3: View Synthesis for Advanced 3D Video

AbstractIn this paper, we present a novel view synthesis algorithm. We are specifically concerned with view interpolation between two reference views. It is known that many holes occur in a virtual view when we synthesize using depth‐image based rendering (DIBR). We introduce the view synthesis algorithm which consists of the adaptive view blending and the directional background pixel replication. Our solution warps the two reference views into the virtual viewpoint and then merges the two projected images. Unlike the early work which blends the uniformly the projected views, we blend the projected images depending on the hole position. Also, to fill up the remaining holes, we develop the background pixel replication which distinguishes between the background and foreground and fills the current occluded pixel with color pixel in the true background. This view synthesis allows for the natural visual quality. We give some results about the subjective quality of the synthesized views and the virtual view produced by our algorithm shows significant visual quality.

Nearest-neighbor and bilinear resampling factor estimation to detect blockiness or blurriness of an image

In digital publishing, a low-resolution image is highly undesirable. Inexperienced users often try to include low-resolution images from the Internet or digital cameras in documents they are composing. Current preflight tools are able to single them out, but what if those low-resolution images have been interpolated? They may have a sufficient resolution, but their quality has been compromised, especially images interpolated by nearest-neighbor (which includes pixel replication) and bilinear interpolation. The interpolated images often display blocky artifacts, blurry artifacts, or loss of texture. We outline novel nearest-neighbor and bilinear interpolation detection algorithms that are designed to estimate rational resampling factors (above 1) in both the vertical and horizontal dimensions. The robustness of these algorithms to several common postprocessing algorithms is also evaluated.

High Quality Raster Image Resizing Using Linear Weighted Manipulation

Due to the versatile customer applications and especially for large size industrial printing needs, the conflict comes between the image enlargement and high resolution quality. The image resizing process introduces specific pixels to the original image, or leaves out specific pixels from the original image. In order to maintain image quality and non-distortion, the value of such pixels and their positions were carefully deliberated. Prior researches such as DCT-based [2], wavelets, pixel replication linear interpolation [1] [3] those were limited by bandwidth and storage spaces, therefore a high efficiency resizing method is proposed in this paper. We introduce a new simple and high quality image-resizing algorithm that using linear weighted method for enlarging and drawing out particular pixels for reducing image size. The enlarging is first procedure and following with reducing procedure, where the image boundaries between those procedures needed to be carefully identification by linear weighted to avoid resolution loss. The simulation suggested the image can be enlarged up to four times and reduced down to quarter of original image at no resolution deterioration, and save the storage space need of fifty percent. Besides, the separation procedure is independent on enlarging procedure that can be operated if the image only needs to be reduced, same performance as mentioned before. This algorithm had been implemented by hardware and it will easily operate in software function in near future.

A locally adaptive zooming algorithm for digital images

Abstract In this paper we address the problem of producing an enlarged picture from a given digital image (zooming). We propose a method that tries to take into account information about discontinuities or sharp luminance variations while doubling the input picture. This is realized by a nonlinear iterative procedure of the zoomed image and could hence be implemented with limited computational resources. The algorithm works on monochromatic images, RGB color pictures and Bayer data images acquired by CCD/CMOS camera sensor. Our experiments show that the proposed method beats in quality classical simple zooming techniques (e.g. pixel replication, simple interpolation). Moreover our algorithm is competitive both for quality and efficiency with bicubic interpolation.

An MRF-Based Approach to Generation of Super-Resolution Images from Blurred Observations

This paper presents a new technique for generating a high resolution image from a blurred image sequences this is also referred to as super-resolution restoration of images. The image sequence consists of decimated, blurred and noisy versions of the high resolution image. The high resolution image is modeled as a Markov random field (MRF) and a maximum a posteriori (MAP) estimation technique is used for super-resolution restoration. Unlike other super-resolution imaging methods, the proposed technique does not require sub-pixel registration of given observations. A simple gradient descent method is used to optimize the functional. The discontinuities in the intensity process can be preserved by introducing suitable line processes. Superiority of this technique to standard methods of image expansion like pixel replication and spline interpolation is illustrated.

A generalized interpolative vector quantization method for jointly optimal quantization, interpolation, and binarization of text images

This paper presents an approach for the effective combination of interpolation with binarization of gray level text images to reconstruct a high resolution binary image from a lower resolution gray level one. We study two nonlinear interpolative techniques for text image interpolation. These nonlinear interpolation methods map quantized low dimensional 2 x 2 image blocks to higher dimensional 4 x 4 (possibly binary) blocks using a table lookup operation. The first method performs interpolation of text images using context-based, nonlinear, interpolative, vector quantization (NLIVQ). This system has a simple training procedure and has performance (for gray-level high resolution images) that is comparable to our more sophisticated generalized interpolative VQ (GIVQ) approach, which is the second method. In it, we jointly optimize the quantizer and interpolator to find matched codebooks for the low and high resolution images. Then, to obtain the binary codebook that incorporates binarization with interpolation, we introduce a binary constrained optimization method using GIVQ. In order to incorporate the nearest neighbor constraint on the quantizer while minimizing the distortion in the interpolated image, a deterministic-annealing-based optimization technique is applied. With a few interpolation examples, we demonstrate the superior performance of this method over the NLIVQ method (especially for binary outputs) and other standard techniques e.g., bilinear interpolation and pixel replication.

Combining multispectral data of differing spatial resolution

An improvement to an existing band correlation method is described for merging high spatial resolution imagery with data from additional lower spatial resolution bands. A figure of merit quantifies the improvement over pixel replication of the lower resolution imagery. Examples are given using AVIRIS data to simulate the future Landsat Thematic Mapper. A physical basis is useful in understanding results of the image sharpening procedure.

Adaptive resampling algorithm for image zooming

Applying an interpolation function indiscriminately to an image, to resample it, will generally result in aliasing, edge blurring and other artefacts. The authors present an adaptive resampling algorithm for zooming up images. The algorithm is based on analysing the local structure of the image and applying a near optimal and least time-consuming resampling function that will preserve edge locations and their contrast. This is done by segmenting the image dynamically into homogeneous areas, as it is scanned or received. Based on the location of the pixel to be computed (whether it is within a homogenous area, is on its edge or is an isolated feature), interpolation, extrapolation or pixel replication is chosen. Algorithm performance, from both a quality and a computational complexity aspect, are compared to different methods and functions previously reported in the literature. The advantage of the algorithm is quite apparent at edges and for large zooming factors.

Interpolator for infrared images

The use of focal plane arrays in IR imaging systems is becoming increasingly important, but in the long-wave IR region the number of detector elements in the array is limited by the current state of technology, and this in turn restricts the available spatial resolution or field of view. An image processing scheme is described that is able to interpolate and enhance an image in a unified moving window operation. The image is first expanded to the required size by pixel replication and then processed so that the resulting spectrum approximates that of the original scene. A numerical method has been developed to calculate the interpolator and its use has been demonstrated in a computer simulation.

Browsing High Definition Colour Pictures

AbstractThe authors describe a method of presenting colour pictures which allows users to browse by panning and zooming. The pictures are seen as though on a 512 by 512 by 12 bit framestore, but are defined to 4096 by 4096. The implementation allows the entire picture to be seen in overview, with fine details averaged, or to be zoomed in upon with finer detail progressively revealed by a sequence of twofold linear magnifications. Further zooming, beyond the resolution of the picture data, automatically produces a conventional pixel replication effect. In addition, the picture may be panned vertically or horizontally.