A novel contrast enhancement algorithm based on the layered difference representation of 2D histograms is proposed in this paper. We attempt to enhance image by amplifying gray-level differences between adjacent pixels. To end, we obtain histogram h(k, k + l ) from an input image, which counts pairs pixels with gray-levels and , represent a tree-like structure. Then, formulate constrained optimization problem observation that differences, occurring more frequently should be emphasized output...

In this work, we develop a new motion-compe (MC) interpolation algorithm to enhance the temporal resolution of video sequences. First, propose bilateral motion estimation scheme obtain field an interpolated frame without yielding hole and overlapping problems. Then, partition into several object regions by clustering vectors. We apply variable-size block MC (VS-BMC) boundaries in order reconstruct edge information with higher quality. Finally, use adaptive overlapped (OBMC), which adjusts...

An adaptive rain streak removal algorithm for a single image is proposed in this work. We observe that typical has an elongated elliptical shape with vertical orientation. Thus, we first detect regions by analyzing the rotation angle and aspect ratio of kernel at each pixel location. then perform nonlocal means filtering on detected selecting neighbor pixels their weights adaptively. Experimental results demonstrate removes streaks more efficiently provides higher restored qualities than...

A novel algorithm to remove rain or snow streaks from a video sequence using temporal correlation and low-rank matrix completion is proposed in this paper. Based on the observation that are too small move fast affect optical flow estimation between consecutive frames, we obtain an initial map by subtracting temporally warped frames current frame. Then, decompose into basis vectors based sparse representation, classify those streak ones outliers with support vector machine. We then refine...

A novel algorithm to segment a primary object in video sequence is proposed this work. First, we generate candidate regions for the using both color and motion edges. Second, estimate initial regions, by exploiting recurrence property of object. Third, augment with missing parts or reducing them excluding noisy repeatedly. This augmentation reduction process (ARP) identifies region each frame. Experimental results demonstrate that significantly outperforms state-of-the-art conventional...

An interactive image segmentation algorithm, which accepts user-annotations about a target object and the background, is proposed in this work. We convert into interaction maps by measuring distances of each pixel to annotated locations. Then, we perform forward pass convolutional neural network, outputs an initial map. However, user-annotated locations can be mislabeled result. Therefore, develop backpropagating refinement scheme (BRS), corrects pixels. Experimental results demonstrate that...

We propose a novel video frame interpolation algorithm based on asymmetric bilateral motion estimation (ABME), which synthesizes an intermediate between two input frames. First, we predict symmetric fields to interpolate anchor frame. Second, estimate motions from the Third, use warp frames backward and reconstruct Last, refine frame, develop new synthesis network that generates set of dynamic filters residual using local global information. Experimental results show proposed achieves...

A novel contrast enhancement algorithm based on the layered difference representation is proposed in this work. We first represent gray-level differences at multiple layers a tree-like structure. Then, observation that differences, occurring more frequently input image, should be emphasized output we solve constrained optimization problem to derive transformation function each layer. Finally, aggregate functions all into overall function. Simulation results demonstrate enhances images...

Presents an integrated system for navigation parameter estimation using sequential aerial images, where the parameters represent positional and velocity information of aircraft autonomous navigation. The proposed is composed two parts: relative position absolute estimation. Relative recursively computes current by accumulating displacement estimates extracted from successive images. Simple accumulation values reduces reliability as goes on navigating, resulting in a large error. Therefore,...

A power-constrained contrast-enhancement algorithm for emissive displays based on histogram equalization (HE) is proposed in this paper. We first propose a log-based modification scheme to reduce overstretching artifacts of the conventional HE technique. Then, we develop power-consumption model and formulate an objective function that consists histogram-equalizing term power term. By minimizing convex optimization theory, achieves contrast enhancement saving simultaneously. Moreover, extend...

We propose a deep learning algorithm for single-image depth estimation based on the Fourier frequency domain analysis. First, we develop convolutional neural network structure and new loss function, called depth-balanced Euclidean loss, to train reliably wide range of depths. Then, generate multiple map candidates by cropping input images with various ratios. In general, cropped image small ratio yields details more faithfully, while that large provides overall distribution reliably. To take...

A novel saliency detection algorithm for video sequences based on the random walk with restart (RWR) is proposed in this paper. We adopt RWR to detect spatially and temporally salient regions. More specifically, we first find a temporal distribution using features of motion distinctiveness, consistency, abrupt change. Among them, distinctiveness derived by comparing profiles image patches. Then, employ as restarting walker. In addition, design transition probability matrix walker spatial...

A semi-supervised online video object segmentation algorithm, which accepts user annotations about a target at the first frame, is proposed in this work. We propagate labels previous frame to current using optical flow vectors. However, propagation error-prone. Therefore, we develop convolutional trident network (CTN), has three decoding branches: separative, definite foreground, and background decoders. Then, perform Markov random field optimization based on outputs of sequentially carry...

We propose a novel algorithm for monocular depth estimation using relative maps. First, convolutional neural network, we estimate depths between pairs of regions, as well ordinary depths, at various scales. Second, restore maps from selectively estimated data based on the rank-1 property pairwise comparison matrices. Third, decompose and into components recombine them optimally to reconstruct final map. Experimental results show that proposed provides state-of-art performance.

A simple yet effective object descriptor for visual tracking is proposed in this paper. We first decompose the bounding box of a target into multiple patches, which are described by color and gradient histograms. Then, we concatenate features spatially ordered patches to represent appearance. Moreover, alleviate impacts background information possibly included box, determine patch weights using random walk with restart (RWR) simulations. The importance each description foreground...

The aim of this paper is to extend the concept cubic sets neutrosophic sets. notions truth-internal (indeterminacy-internal, falsity-internal) and truth-external (indeterminacy-external, falsity-external) are introduced, related properties investigated.

Recently, the encoder-decoder and intensity transformation approaches lead to impressive progress in image enhancement. However, often loses details input images during down-sampling up-sampling processes. Also, has a limited capacity cover color between low-quality high-quality images. In this paper, we propose novel approach, called representative transform (RCT), tackle these issues existing methods. RCT determines different colors specialized estimates transformed for colors. It then...

A novel light field super-resolution algorithm to improve the spatial and angular resolutions of images is proposed in this work. We develop (SR) networks, which can faithfully interpolate domains regardless coordinates. For each input image, we feed adjacent into SR networks extract multi-view features using a trainable disparity estimator. concatenate remix them through adaptive feature remixing (AFR) module, performs channel-wise pooling. Finally, remixed used augment or resolution....

A novel ordinal regression algorithm, called moving window (MWR), is proposed in this paper. First, we propose the notion of relative rank ( <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\rho$</tex> -rank), which a new order representation scheme for input and reference instances. Second, develop global local regressors -regressors) to predict -ranks within entire specific ranges, respectively. Third, refine an initial estimate iteratively by...

We propose a novel postprocessing technique, based on the theory of projections onto convex sets (POCS), to reduce blocking artifacts in transform coded images. It is assumed, our approach, that original image highly correlated. Thus, global frequency characteristics two adjacent blocks are similar local ones each block. consider high-frequency components decoded image, which not found ones, as results from artifact. employ an N-point discrete cosine (DCT) obtain characteristics, and...

Two novel error concealment techniques are proposed for video transmission over noisy channels in this work. First, we present a spatial method to compensate lost macroblock intra-coded frames, which no useful temporal information is available. Based on selective directional interpolation, our can recover both smooth and edge areas efficiently. Second, examine dynamic mode-weighted replenishing missing pixels of inter-coded frames. Our adopts decoder-based tracking model combines several...