Image specific discriminative feature extraction for skin segmentation

In recent times, the majority of colour-based skin detection methods used skin modelling in different colour spaces, and they are capable of skin classification at a pixel level. However, the accuracy of these methods is significantly affected by different issues, such as the presence of skin-like colours in scene background, variations in skin pigmentation, scene illumination, etc. Recent developments show that the discriminating power of a colour-based skin classifier can be increased by employing texture and spatial features. However, we observed that discriminability between skin and non-skin regions does not follow any statistics, and the discrimination is extremely image specific. In this paper, a novel adaptive discriminative analysis (ADA) is proposed to extract most discriminant features between skin and non-skin regions from an image itself in an unsupervised manner. Experimental results for standard databases show that the proposed method can efficiently segment out skin pixels in the presence of skin-like background colours.

     

Quality-distinguishing and patch-comparing no-reference image quality assessment

No-reference image quality assessment (NR-IQA) based on deep learning attracts a great research attention recently. However, its performance in terms of accuracy and efficiency is still under exploring. To address these issues, in this paper, we propose a quality-distinguishing and patch-comparing NR-IQA approach based on convolutional neural network (QDPC-CNN). We improve the prediction accuracy by two proposed mechanisms: quality-distinguishing adaption and patch-comparing regression. The former trains multiple models from different subsets of a dataset and adaptively selects one for predicting quality score of a test image according to its quality level, and the latter generates patch pairs for regression under different combination strategies to make better use of reference images in network training and enlarge training data at the same time. We further improve the efficiency of network training by a new patch sampling way based on the visual importance of each patch. We conduct extensive experiments on several public databases and compare our proposed QDPC-CNN with existing state-of-the-art methods. The experimental results demonstrate that our proposed method outperforms the others both in terms of accuracy and efficiency.

     

A salt and pepper noise image denoising method based on the generative classification

In this paper, an image denoising algorithm is proposed for salt and pepper noise. First, a generative model is built on a patch as a basic unit and then the algorithm locates the image noise within that patch in order to better describe the patch and obtain better subsequent clustering. Second, the algorithm classifies patches using a generative clustering method, which provides additional similarity information for noise repairing, suppresses the interference of noise and abandons those classes that consist of a smaller number of patches. Finally, the algorithm builds a non-local switching filter to remove the salt and pepper noise. Simulation results show that the proposed algorithm effectively denoises salt and pepper noise of various densities. It obtains a better visual quality and higher peak signal-to-noise ratio score than several state-of-the-art algorithms. In short, our algorithm uses a noisy patch as the basic unit, a patch clustering method to optimize the repair data set as well as obtains a better denoising effect, and provides a guideline for future denoising and repair methods.

     

Low-complexity two-step lossless depth coding using coarse Lossy coding

Texture and depth images are generally used for 3D viewing with advanced displays. Because sthe characteristics of a depth image are very different from those of a texture image, an efficient compression method is required to transmit a depth image in a limited bandwidth. In this paper, a low-complexity two-step lossless depth coding (LTLDC) method using coarse lossy coding is proposed. The proposed method downsamples an original image and then coarsely compresses the downsampled image in the first step. This compressed image is upsampled, and then its residual is generated by subtracting the upsampled image from the original image. In the second step, each coding block within the residual and original images is adaptively compressed with a fast mode decision method in a lossless way, and the proposed method determines the best block based on their coding performance. Experimental results show that the proposed LTLDC method achieves a bitrate reduction of 4.35% with encoding complexity reduction of 20.38%.

     

Denoising Deep Monte Carlo Renderings

We present a novel algorithm to denoise deep Monte Carlo renderings, in which pixels contain multiple colour values, each for a different range of depths. Deep images are a more expressive representation of the scene than conventional flat images. However, since each depth bin receives only a fraction of the flat pixel's samples, denoising the bins is harder due to the less accurate mean and variance estimates. Furthermore, deep images lack a regular structure in depth—the number of depth bins and their depth ranges vary across pixels. This prevents a straightforward application of patch‐based distance metrics frequently used to improve the robustness of existing denoising filters. We address these constraints by combining a flat image‐space non‐local means filter operating on pixel colours with a deep cross‐bilateral filter operating on auxiliary features (albedo, normal, etc.). Our approach significantly reduces noise in deep images while preserving their structure. To our best knowledge, our algorithm is the first to enable efficient deep‐compositing workflows with denoised Monte Carlo renderings. We demonstrate the performance of our filter on a range of scenes highlighting the challenges and advantages of denoising deep images.     

Colour Names as Response Criteria∗

Abstract

A study was conducted to determine the consistency of colour naming of chemical reaction spots among subjects and trained chemists. The following conclusions were drawn.

1. Single, arbitrarily assigned names are inadequate to convey unequivocal meaning to a group of observers who must base important decisions on their judgment of colour.

2. When it was logically possible to group a large number of names into one or two overall categories, observer agreement is markedly increased. 3. The study of such regrouped names together with the associated colours makes it possible to construct a colour continuum-bar which obviates the need for colour naming altogether, but facilitates direct comparison of spot colours with criterion colours.     

Region duplication detection based on image segmentation and keypoint contexts

In this paper, a novel region duplication detection method is proposed based on image segmentation and keypoint contexts. The proposed method includes the primary region duplication detection based on keypoints and the supplementary region duplication detection based on blocks. In the primary region duplication detection, an image is divided into non-overlapped patches by using SLIC. Furthermore, the keypoints are matched and clustered within the same patch as patch feature. Then the patches are matched and an affine transformation is tried to be estimated from a pair of patches. When the estimation fails, in the supplementary region duplication detection, a transformation matrix is tried to be estimated from a pair of keypoints by the proposed Keypoint Contexts (KC) approach. The experimental results indicate that the proposed method can achieve much better comprehensive performances than the state-of-the-art methods on the public databases, even under various challenging conditions.

     

Adaptive colour constancy algorithm using discrete wavelet transform

The colours of chromatically homogeneous object surfaces measured by a sensor vary with the illuminant colour used to illuminate the objects. In contrast, colour constancy enables humans to identify the true colours of the surfaces under varying illumination. This paper proposes an adaptive colour constancy algorithm which estimates the illuminant colour from wavelet coefficients at each scale of the decomposition by discrete wavelet transform of the input image. The angular error between the estimated illuminant colours in consecutive scales are used to determine the optimum scale for the best estimate of the true illuminant colour. The estimated illuminant colour is then used to modify the approximation subbands of the image so as to generate the illuminant-colour corrected image via inverse discrete wavelet transform. The experiments show that the colour constancy results generated by the proposed algorithm are comparable or better than those of the state-of-the-art colour constancy algorithms that use low-level image features.     

Improving the selection of feature points for tracking

The problem considered in this paper is how to select the feature points (in practice, small image patches are used) in an image from an image sequence, such that they can be tracked adequately further through the sequence. Usually, the tracking is performed by some sort of local search method looking for a similar patch in the next image in the sequence. Therefore, it would be useful if we could estimate the size of the convergence region for each image patch. There is a smaller chance of error when calculating the displacement for an image patch with a large convergence region than for an image patch with a small convergence region. Consequently, the size of the convergence region can be used as a proper goodness measure for a feature point. For the standard Kanade-Lucas-Tomasi (KLT) tracking method, we propose a simple and fast way to approximate the convergence region for an image patch. In the experimental part, we test our hypothesis on a large set of real data.     

An image fusion framework using morphology and sparse representation

Image fusion is the process which aims to integrate the relevant and complementary information from a set of images into a single comprehensive image. Sparse representation (SR) is a powerful technique used in a wide variety of applications like denoising, compression and fusion. Building a compact and informative dictionary is the principal challenge in these applications. Hence, we propose a supervised classification based learning technique for the fusion algorithm. As an initial step, each patch of the training data set is pre-classified based on their gradient dominant direction. Then, a dictionary is learned using K-SVD algorithm. With this universal dictionary, sparse coefficients are estimated using greedy OMP algorithm to represent the given set of source images in the dominant direction. Finally, the Euclidean norm is used as a distance measure to reconstruct the fused image. Experimental results on different types of source images demonstrate the effectiveness of the proposed algorithm with conventional methods in terms of visual and quantitative evaluations.

     

Ensemble dictionary learning for saliency detection

The human visual system (HSV) is quite adept at swiftly detecting objects of interest in complex visual scene. Simulating human visual system to detect visually salient regions of an image has been one of the active topics in computer vision. Inspired by random sampling based bagging ensemble learning method, an ensemble dictionary learning (EDL) framework for saliency detection is proposed in this paper. Instead of learning a universal dictionary requiring a large number of training samples to be collected from natural images, multiple over-complete dictionaries are independently learned with a small portion of randomly selected samples from the input image itself, resulting in more flexible multiple sparse representations for each of the image patches. To boost the distinctness of salient patch from background region, we present a reconstruction residual based method for dictionary atom reduction. Meanwhile, with the obtained multiple probabilistic saliency responses for each of the patches, the combination of them is finally carried out from the probabilistic perspective to achieve better predictive performance on saliency region. Experimental results on several open test datasets and some natural images demonstrate that the proposed EDL for saliency detection is much more competitive compared with some existing state-of-the-art algorithms.     

Selecting colour codes for a computer-generated topographic map based on perception experiments and functional requirements

Abstract

This paper describes a series of human factors analyses that guided the selection of chromaticities and luminances for a computer-generated topographic map. By virtue of its impressive computational capabilities, this CRT-displayed digital map will greatly facilitate the navigational accuracy and situational awareness of army helicopter aviators during low level and nap-of-the-earth flight. Colour codes were assigned to the digital map's point, linear and area features according to guidelines derived from four colour naming and two symbol search experiments. The design of each study was structured around the map's functional requirements: the five linear feature colours should have high luminance and support absolute colour identification; the three point symbol colours should be identifiable at small sizes; and the four area colours should minimize colour distortions, with the two terrain colours luminance-shaded to depict elevation information. Within these constraints, the results of the colour naming studies yielded an initial set of map colour codes by identifying the most frequently occurring colour confusions arising from the perceptual distortions of brightness contrast, colour contrast and Gaussian spread. The symbol search studies further refined colour selection by identifying the specific foreground/background colour combinations that hinder search, and by quantifying the conditions under which a colour or monochrome map facilitates symbol search.     

Saliency-based deep convolutional neural network for no-reference image quality assessment

In this paper, we proposed a novel method for No-Reference Image Quality Assessment (NR-IQA) by combining deep Convolutional Neural Network (CNN) with saliency map. We first investigate the effect of depth of CNNs for NR-IQA by comparing our proposed ten-layer Deep CNN (DCNN) for NR-IQA with the state-of-the-art CNN architecture proposed by Kang et al. (2014). Our results show that the DCNN architecture can deliver a higher accuracy on the LIVE dataset. To mimic human vision, we introduce saliency maps combining with CNN to propose a Saliency-based DCNN (SDCNN) framework for NR-IQA. We compute a saliency map for each image and both the map and the image are split into small patches. Each image patch is assigned with a patch importance value based on its saliency patch. A set of Salient Image Patches (SIPs) are selected according to their saliency and we only apply the model on those SIPs to predict the quality score for the whole image. Our experimental results show that the SDCNN framework is superior to other state-of-the-art approaches on the widely used LIVE dataset. The TID2008 and the CISQ image quality datasets are utilised to report cross-dataset results. The results indicate that our proposed SDCNN can generalise well on other datasets.     

Bayesian Scheme for Interactive Colourization,Recolourization and Image/Video Editing

We propose a general image and video editing method based on a Bayesian segmentation framework. In the first stage, classes are established from scribbles made by a user on the image. These scribbles can be considered as a multi‐map (multi‐label map) that defines the boundary conditions of a probability measure field to be computed for each pixel. In the second stage, the global minima of a positive definite quadratic cost function with linear constraints, is calculated to find the probability measure field. The components of such a probability measure field express the degree of each pixel belonging to spatially smooth classes. Finally, the computed probabilities (memberships) are used for defining the weights of a linear combination of user provided colours or effects associated to each class. The proposed method allows the application of different operators, selected interactively by the user, over part or the whole image without needing to recompute the memberships. We present applications to colourization, recolourization, editing and photomontage tasks.     

Metamodelling of wheel–rail normal contact in railway crossings with elasto-plastic material behaviour

A metamodel considering material plasticity is presented for computationally efficient prediction of wheel–rail normal contact in railway switches and crossings (S&C). The metamodel is inspired by the contact theory of Hertz, and for a given material, it computes the size of the contact patch and the maximum contact pressure as a function of the normal force and the local curvatures of the bodies in contact. The model is calibrated based on finite element (FE) simulations with an elasto-plastic material model and is demonstrated for rail steel grade R350HT. The error of simplifying the contact geometry is discussed and quantified. For a moderate difference in contact curvature between wheel and rail, the metamodel is able to accurately predict the size of the contact patch and the maximum contact pressure. The accuracy is worse when there is a small difference in contact curvature, where the influence of variation in curvature within the contact patch becomes more significant. However, it is shown that such conditions lead to contact stresses that contribute less to accumulated plastic deformation. The metamodel allows for a vast reduction of computational effort compared to the original FE model as it is given in analytical form.

     

Adaptive speckle reduction using combined multiresolution analysis and prediction technique

Speckle appearing in SAR images as granular noise, must be reduced for good image classification and interpretation. In this paper, two techniques for speckle reduction are presented: an adaptive prediction technique and a multiresolution thresholding technique. When taking into account the tradeoff between speckle reduction and texture preservation, some limitations in terms of speckle reduction rate are observed. Two efficient schemes based on combinations of the two separate techniques are presented, and are shown to give better performance than the separate schemes in terms of speckle reduction and texture preservation. This is made possible because of the specific advantages of each technique.     

Implementation for noise suppression in images

Random noise in images represents the primary problem for early visual processing. This paper describes an adaptive surface labelling technique (ASL) that suppresses image noise by using data-driven rules that concern surface continuity. The algorithm first obtains a global estimate of the noise distribution and then tries to fit a surface to each part of the image. If this can be done so that the error in fitting a surface is within what would be expected from known noise statistics, then the central pixel is reset to lie on that surface. Analytical results are presented which demonstrate algorithm success provided that the number of samples in the patch is greater than 2m, where m is the number of linear parameters that determine the form of the patch.     

Salient object detection with low-rank approximation and ℓ2,1-norm minimization

Salient object detection is an important issue in computer vision and image procession in that it can facilitate humans to locate conspicuous visual regions in complex scenes rapidly and improve the performance of object detection and video tracking. In recent years, low-rank matrix approximation has been proved to be favorable in image saliency detection and gained a great deal of attention. An underlying assumption of low-rank recovery is that an image is a combination of background regions being low-rank and salient objects being sparse, which corresponds to tough non-smooth optimization problems. In this paper, by incorporating ℓ_2,1-norm minimization, we obtain the corresponding smooth optimization problems and propose two effective algorithms with proved convergence. To guarantee the robustness of the proposed methods, the input image is divided into patches and each patch is approximately represented by its mean value. Besides, multi-scale visual features of each patch of the given image are extracted to capture common low-level features such as color, edge, shape and texture. The salient objects of a given image are indicated with sparse coefficients solved by the low-rank matrix approximation problem. Saliency maps are further produced with integration of the high-level prior knowledge. Finally, extensive experiments in four real-world datasets demonstrate that the proposed methods come with competitive performance over the eight compared state-of-the-arts.     

结合支持向量回归和图像自相似的单幅图像超分辨率算法

目的 基于学习的单幅图像超分辨率算法是借助实例训练库由一幅低分辨率图像产生高分辨率图像。提出一种基于图像块自相似性和对非线性映射拟合较好的支持向量回归模型的单幅超分辨率方法,该方法不需使用外部图像训练库。方法 首先根据输入的低分辨率图像建立图像金字塔及包含低/高分辨率图像块对的集合;然后在低/高分辨率图像块对的集合中寻找与输入低分辨率图像块的相似块,利用支持向量回归模型学习这些低分辨率相似块和其对应的高分辨率图像块的中心像素之间的映射关系,进而得到未知高分辨率图像块的中心像素。结果 为了验证本文设计算法的有效性,选取结构和纹理不同的7幅彩色高分辨率图像,对其进行高斯模糊的2倍下采样后所得的低分辨率图像进行超分辨率重构,与双三次插值、基于稀疏表示及基于支持向量回归这3个超分辨率方法重建的高分辨率图像进行比较,峰值信噪比平均依次提升了2.37 dB、0.70 dB和0.57 dB。结论 实验结果表明,本文设计的算法能够很好地实现图像的超分辨率重构,特别是对纹理结构相似度高的图像具有更好的重构效果。