Input Color Research Articles

Fuzzy averaging filter for impulse noise reduction in colour images with a correction step

In this paper we propose a fuzzy detection and reduction method for impulse noise in colour images. Detection is based on the fuzzyfication of a well-known statistic called ROD. The noise degrees obtained are used to reduce impulses by employing a fuzzy averaging between the input colour vector and a robust estimate of noise-free colour vector within the input neighbourhood. Fuzzy averaging has some advantages in terms of both noise reduction and detail preservation in front of detect and replace approaches because of threshold based decisions of the latter. However, robustness of the former is lower. We solve this problem by including a correction mechanism that checks the fuzzy noise degree of the output and replaces it with a robust colour vector either when noise has not been properly reduced or when a colour artefact has been introduced. We carry out a thorough study of the method parameter setting and give a convenient and robust setting. Experimental results show that our approach is very robust in front of four different types of impulse noise.

CONTRAST ENHANCEMENT FOR COLOR IMAGES USING AN ADJUSTABLE CONTRAST STRETCHING TECHNIQUE

With the growing demand for high-quality color images, efficient yet low-complexity methods are increasingly needed for better visualization. Unfortunately, the low-contrast is one prevalent effect that degrades color images due to various unavoidable limitations. Hence, a new adjustable contrast stretching technique is proposed in this article to improve the contrast of color images. The processing scheme of the proposed technique is relatively simple. It starts by converting the input color image to grayscale. Then, it automatically computes two contrast tuning parameters depending on the pre-determined grayscale image. Finally, it improves the contrast of the degraded color image using an amended version of an existing contrast stretching technique. Accordingly, its input is a color image and a contrast adjustment parameter δ, while its output is a contrast-adjusted color image. The proposed technique is tested by conducting intensive experiments on real-degraded images, and it is compared with four well-known contrast enhancement techniques. In addition, the proposed and comparative techniques are evaluated based on three eminent no-reference image quality assessment metrics. From the performance analysis of the achieved experiments and comparisons, the proposed technique provided satisfying performances and outperformed the comparative techniques in terms of recorded accuracy and perceived quality.

Color encoding in biologically-inspired convolutional neural networks

Convolutional Neural Networks have been proposed as suitable frameworks to model biological vision. Some of these artificial networks showed representational properties that rival primate performances in object recognition. In this paper we explore how color is encoded in a trained artificial network. It is performed by estimating a color selectivity index for each neuron, which allows us to describe the neuron activity to a color input stimuli. The index allows us to classify whether they are color selective or not and if they are of a single or double color. We have determined that all five convolutional layers of the network have a large number of color selective neurons. Color opponency clearly emerges in the first layer, presenting 4 main axes (Black-White, Red-Cyan, Blue-Yellow and Magenta-Green), but this is reduced and rotated as we go deeper into the network. In layer 2 we find a denser hue sampling of color neurons and opponency is reduced almost to one new main axis, the Bluish-Orangish coinciding with the dataset bias. In layers 3, 4 and 5 color neurons are similar amongst themselves, presenting different type of neurons that detect specific colored objects (e.g., orangish faces), specific surrounds (e.g., blue sky) or specific colored or contrasted object-surround configurations (e.g. blue blob in a green surround). Overall, our work concludes that color and shape representation are successively entangled through all the layers of the studied network, revealing certain parallelisms with the reported evidences in primate brains that can provide useful insight into intermediate hierarchical spatio-chromatic representations.

Robust Color Image Hashing Using Quaternion Polar Complex Exponential Transform for Image Authentication

Image hashing is one of the multimedia protection techniques. In this paper, a new method for robust image hashing based on quaternion polar complex exponential transform (QPCET) is proposed. The proposed method targets two goals. The first goal is the robustness against geometric and common signal processing attacks. The second one is authenticating color images without conversion which keeps their color information. In the proposed method, the input color image is normalized by the bicubic interpolation and then the interpolated image passes to Gaussian low-pass filter. QPCET moments are used to extract features. Finally, the hash value is calculated using the extracted features. On the sender side, a secret key is utilized to increase the protection of the hash value before transmitting it. The hash value is attached with the transmitted color image. On the receiver side, the authenticity of the received image is checked by decrypting the hash value. Euclidean distance is used to check the similarity between different hashes. Results of the conducted experiments prove the robustness of proposed hash against different geometric and signal processing attacks. Also, it preserves the content of the transmitted color image. Hashing different images has a very low collision probability which ensure the suitability of the proposed method for image authentication. Comparison with the existing methods ensures the superiority of the proposed method.

Ocularity Feature Contrast Attracts Attention Exogenously.

An eye-of-origin singleton, e.g., a bar shown to the left eye among many other bars shown to the right eye, can capture attention and gaze exogenously or reflexively, even when it appears identical to other visual input items in the scene and when the eye-of-origin feature is irrelevant to the observer’s task. Defining saliency as the strength of exogenous attraction to attention, we say that this eye-of-origin singleton, or its visual location, is salient. Defining the ocularity of a visual input item as the relative difference between its left-eye input and its right-eye input, this paper shows the general case that an ocularity singleton is also salient. For example, a binocular input item among monocular input items is salient, so is a left-eye-dominant input item (e.g., a bar with a higher input contrast to the left eye than to the right eye) among right-eye-dominant items. Saliency by unique input ocularity is analogous to saliency by unique input colour (e.g., a red item among green ones), as colour is determined by the relative difference(s) between visual inputs to different photoreceptor cones. Just as a smaller colour difference between a colour singleton and background items makes this singleton less salient, so does a smaller ocularity difference between an ocularity singleton and background items. While a salient colour difference is highly visible, a salient ocularity difference is often perceptually invisible in some cases and discouraging gaze shifts towards it in other cases, making its behavioural manifestation not as apparent. Saliency by ocularity contrast provides another support to the idea that the primary visual cortex creates a bottom-up saliency map to guide attention exogenously.

Optimal Color Multiplexing for Low-Cost Structured Light 3D Capture System with Two Projectors

3D shape reconstruction is one of the most important topics in computer vision and the foundation for a wide field of application. Among various technologies, structured light is one of the most reliable techniques. However, given the field of view of projectors and cameras available in the market, the working distance needed for projectors is typically larger than that for cameras. To reduce the working distance of the projectors while covering the whole working platform, two projectors with their field of view overlapping are used to cover the working area which holds objects to be scanned. We present a spectral analysis based model for the projector-camera system, in order to find the most distinguishable colors for two projectors, and best separate the projected patterns from two projectors. The optimal values of the two colors are determined by the pattern search method in the presence of noise, which is modeled as multivariate Gaussian noise, and characterized for different input colors. The camera sensors' responses to the projector are measured after linearization with gray balance curves. After being properly calibrated, based on one image shot of the object with binary M-array patterns projected on it, the system can reconstruct a 3D shape of the object surface.

Effect of Decolorized Images In Scene Classification Using Deep Convolution Features

Scene classification is considered as an imperative issue for computer vision and has got extensive consideration in the recent past. Due to recent developments in high performance computing units such as GPUs, popularly known deep learning algorithm namely, convolutional neural networks (CNNs), exploits huge datasets to give powerful models. The paper proposes the use of transfer learning technique, by which a pre-trained model known as Places-CNN is used to generate feature vectors for each scene image of the dataset. The scene-classification experiments are conducted on the Oliva Torralba (OT) scene dataset, which consists of eight outdoor scene categories. The features were extracted from the fully connected layer of the pre-trained Places CNN architecture. The deep features were extracted from the input color images and the grayscale images converted using two different techniques based on singular value decomposition (SVD). The results obtained from classification experiments show that, models trained on SVD-Decolorized and Modified-SVD decolorized images give comparable performance to the input color images. Unlike the color images, which use three planes (RGB) of information, the grayscale images use only one plane of information. The grayscale images were able to retain the required shape and texture information from the original RGB images and, thus sufficient to categorize the classes of scene images.

Image Super-Resolution Reconstruction Based on Disparity Map and CNN

This paper presents a new method for registration of digital micro-mirror device (DMD) cameras through image reconstruction. The method called dual-channel super-resolution reconstruction (DCSR) uses a dual-channel input convolutional neural network method to iteratively obtain high resolution images. The originality of this study relies on the input color weight disparity map that enables extracting edges cues and feature cues effectively during the super-resolution. The iterative initial disparity map is obtained by a cost aggregation method based on the color weights, and the model parameters are updated during the iterative process. To validate performance, a series of the comparison experiments is developed on different network structures and parameter settings for the diverse deconv-pooling layers, convolution kernel sizes, and activation function. Experimental results show that the DCSR method can solve the sub-pixel registration problem of DMD cameras and achieve a good image reconstruction effect. The DCSR performs higher accuracy and an efficient process compared with traditional reconstruction made by single-channel input convolution neural network. The spatial resolution of the reconstructed image is significantly improved, and the detail resolution is enhanced.

WpmDecolor: weighted projection maximum solver for contrast-preserving decolorization

This paper presents a novel semi-reference inspired color-to-gray conversion model for faithfully preserving the contrast details of the color image, essentially differs from most of the no-reference and reference approaches. In the proposed model, on the basic assumption that a good gray conversion should make the conveyed gradient values (i.e., contrast) to be maximal, we present a projection maximum function to model the decolorization procedure. We further incorporate weights of the original gradients into the maximum function. The Gaussian weighted factor consisting of the gradients of each channel of the input color image is employed to better reflect the degree of preserving feature discriminability and color ordering in color-to-gray conversion. The projected gradient descent and discrete searching techniques are developed to solve the proposed model with and without nonnegative constraint, respectively. Extensive experimental evaluations on two existing datasets, containing abundant colors and patterns, show that the proposed method outperforms the state-of-the-art methods quantitatively and qualitatively.

Automatic color image contrast enhancement using Gaussian mixture modeling, piecewise linear transformation, and monotone piecewise cubic interpolant

Conventional contrast enhancement algorithms use complicated nonlinear mapping functions with parameters specified manually. In this paper, a parameter-free piecewise linear transformation and monotone piecewise cubic interpolation are used for color image contrast enhancement. The Gaussian mixture model is used to divide luminance histogram of input color image into multiple sub-histograms. Each sub-histogram is mapped linearly or nonlinearly to a portion of available output dynamic range proportional to its extent and the number of its pixels. The proposed algorithm has more controlled contrast enhancement than conventional algorithms to preserve natural outlook and local details in the input color image. This is done by considering brightness histogram preservation of input image by proposed algorithm, due to considering extent of each sub-histogram for allocating a portion of output dynamic range to it. Experimental results show that the proposed algorithm produces better or comparable contrast-enhanced images with natural outlook than several conventional algorithms on various types of images.

An adaptive RGB colour enhancement formulation for logarithmic image processing-based algorithms

This paper presents an effective colour enhancement framework for statistical and logarithmic image processing (LIP)-based enhancement algorithms. The proposed approach utilizes the fusion of partial, multiple computed luminance channels with colour image channel statistics obtained from the input colour image for adaptive colour enhancement. The proposed scheme does not modify the image intensity channel, avoiding colour fading typically observed in colour images processed with conventional algorithms. The colour enhancement scheme compensates for the weaknesses of greyscale-based contrast enhancement and illumination normalization algorithms by focusing on preserving/restoring or enhancing colour. The proposed system avoids the conversion to complex, non-linear colour spaces such as HSI and HSV while producing similar results without manual adjustment of parameters. Additionally, an adaptive scheme for detection of images with unbalanced colour and uneven illumination is combined with the proposed system. Results show that the proposed scheme augments colour results of greyscale-based contrast enhancement algorithms and is relatively less complex compared to most algorithms in the literature.

Segmentation and semantic labelling of RGBD data with convolutional neural networks and surface fitting

We present an approach for segmentation and semantic labelling of RGBD data exploiting together geometrical cues and deep learning techniques. An initial over‐segmentation is performed using spectral clustering and a set of non‐uniform rational B‐spline surfaces is fitted on the extracted segments. Then a convolutional neural network (CNN) receives in input colour and geometry data together with surface fitting parameters. The network is made of nine convolutional stages followed by a softmax classifier and produces a vector of descriptors for each sample. In the next step, an iterative merging algorithm recombines the output of the over‐segmentation into larger regions matching the various elements of the scene. The couples of adjacent segments with higher similarity according to the CNN features are candidate to be merged and the surface fitting accuracy is used to detect which couples of segments belong to the same surface. Finally, a set of labelled segments is obtained by combining the segmentation output with the descriptors from the CNN. Experimental results show how the proposed approach outperforms state‐of‐the‐art methods and provides an accurate segmentation and labelling.

Image coding based on classified vector quantisation using edge orientation patterns

Vector quantisation (VQ) shows a good performance for image coding with high-compression ratios. However, there are many difficulties for image coding with VQ, especially the edge degradation and high-computational complexity. To resolve these two problems, the authors propose a new coding method based on edge orientation patterns (EOPs) by classifying image blocks into nine classes according to their edge orientations. For colour image coding, 27 codebooks (nine for each colour component) are pre-designed based on a series training images. In the encoding stage, an input colour image is decomposed into Y, Cb, and Cr components, and each component image is divided into non-overlapping 4 × 4 blocks. For each block, eight edge orientation templates of size 4 × 4 are performed to determine its edge orientation. According to the edge orientation, each block is compressed by using the corresponding codebook. Essentially, the authors’ scheme is a kind of classified VC (CVQ). Simulation results show that, their EOP-based CVQ can largely improve the compression efficiency as well as speeding up the encoding process and it is sufficient to establish effectiveness of the authors’ algorithm as compared with the existing techniques.

PERANCANGAN APLIKASI PENYISIPAN CITRA BERWARNA PADA QR CODE DENGAN METODE HALFTONE

The project is made in the form of a software application that accepts input color image to be processed into Halftone QR-Code. Halftone QR-Code is a composite image of the QR-Code and halftone image that can be recognized by QR-Code Scanner. QR-Code image generated using the Library Massagingtoolkit-QRCode with alphanumeric text input degan various versions QR-Code in accordance with a lot of character input. The input image is processed into a halftone image with methods Ordered Dithering variation dither matrix 4 x 4 and 8 x 8. The image of the QR-Code and halftone images will be processed into QR-Code Halftone Pattern Assignment method. Output generated by the application Halftone QR-Code tested with QR-Code ZXing Library and QRAFTER application on Iphone 6s Plus and QR. The average rate of overall success with the test scenario ZXing QR-Code Library by 94%. The average rate of overall success with the test scenario QRAFTER application on Iphone 6s Plus from the laptop screen and hardcopy respectively by 100%.

Log-Euclidean Metrics for Contrast Preserving Decolorization.

This paper presents a novel Log-Euclidean metric inspired color-to-gray conversion model for faithfully preserving the contrast details of color image, which differs from the traditional Euclidean metric approaches. In the proposed model, motivated by the fact that Log-Euclidean metric has promising invariance properties such as inversion invariant and similarity invariant, we present a Log-Euclidean metric-based maximum function to model the decolorization procedure. The Gaussian-like penalty function consisting of the Log-Euclidean metric between gradients of the input color image and transformed grayscale image is incorporated to better reflect the degree of preserving feature discriminability and color ordering in color-to-gray conversion. A discrete searching algorithm is employed to solve the proposed model with linear parametric and non-negative constraints. Extensive evaluation experiments show that the proposed method outperforms the state-of-the-art methods both quantitatively and qualitatively.

Plasmonic nanohole electrodes for active color tunable liquid crystal transmissive pixels.

Plasmonic pixels have been shown to offer numerous advantages over pigment-based color filters used in modern commercial liquid crystal (LC) displays. However, wideband dynamic tunability across the visible spectrum remains challenging. We experimentally demonstrate transmissive electrically tunable LC-nanohole pixels operating across the visible spectrum with unpolarized input light. An ultrathin Al nanohole electrode is designed to exhibit a polarized spectral response based on surface plasmon resonances. An output analyzer in combination with a nematic LC layer enables pixel color to be electronically controlled through an applied voltage across the device, where LC reorientation leads to tunable mixing of the relative contributions from the plasmonic color input. The nanostructured Al layer, acting as a combined electrode, polarizer, and functional color filter, is highly promising for electro-optic display applications.

Inverse Diffusion Curves Using Shape Optimization.

The inverse diffusion curve problem focuses on automatic creation of diffusion curve images that resemble user provided color fields. This problem is challenging since the 1D curves have a nonlinear and global impact on resulting color fields via a partial differential equation (PDE). We introduce a new approach complementary to previous methods by optimizing curve geometry. In particular, we propose a novel iterative algorithm based on the theory of shape derivatives. The resulting diffusion curves are clean and well-shaped, and the final image closely approximates the input. Our method provides a user-controlled parameter to regularize curve complexity, and generalizes to handle input color fields represented in a variety of formats.

Malignancy Detection on Mammography Using Dual Deep Convolutional Neural Networks and Genetically Discovered False Color Input Enhancement.

Breast cancer is the most prevalent malignancy in the US and the third highest cause of cancer-related mortality worldwide. Regular mammography screening has been attributed with doubling the rate of early cancer detection over the past three decades, yet estimates of mammographic accuracy in the hands of experienced radiologists remain suboptimal with sensitivity ranging from 62 to 87% and specificity from 75 to 91%. Advances in machine learning (ML) in recent years have demonstrated capabilities of image analysis which often surpass those of human observers. Here we present two novel techniques to address inherent challenges in the application of ML to the domain of mammography. We describe the use of genetic search of image enhancement methods, leading us to the use of a novel form of false color enhancement through contrast limited adaptive histogram equalization (CLAHE), as a method to optimize mammographic feature representation. We also utilize dual deep convolutional neural networks at different scales, for classification of full mammogram images and derivative patches combined with a random forest gating network as a novel architectural solution capable of discerning malignancy with a specificity of 0.91 and a specificity of 0.80. To our knowledge, this represents the first automatic stand-alone mammography malignancy detection algorithm with sensitivity and specificity performance similar to that of expert radiologists.

Two Novel Arctangent Normalized Subband Adaptive Filter Algorithms Against Impulsive Interferences

In this paper, two arctangent-based normalized subband adaptive filter (Arc-NSAF) algorithms are proposed, which are derived from two different forms of arctangent cost function and their difference is the order of the arctangent computation and the summation operation of the normalized subband error signal. Benefiting from the excellent property of the arctangent function for suppressing impulsive interferences, the proposed Arc-NSAF algorithms not only possess good robustness against impulsive interferences, but also obtain faster convergence rate for the colored input signals. To further improve the convergence performance of the Arc-NSAFs when identifying sparse system, the proportionate versions of the Arc-NSAF algorithms are also proposed. Simulation results have demonstrated the superiority of the proposed algorithms in impulsive interferences environments as compared to their respective competitors.

Performance Analysis of Segmentation of Hyperspectral Images Based on Color Image Segmentation

Image segmentation is a fundamental approach in the field of image processing and based on user’s application .This paper propose an original and simple segmentation strategy based on the EM approach that resolves many informatics problems about hyperspectral images which are observed by airborne sensors. In a first step, to simplify the input color textured image into a color image without texture. The final segmentation is simply achieved by a spatially color segmentation using feature vector with the set of color values contained around the pixel to be classified with some mathematical equations. The spatial constraint allows taking into account the inherent spatial relationships of any image and its color. This approach provides effective PSNR for the segmented image. These results have the better performance as the segmented images are compared with Watershed & Region Growing Algorithm and provide effective segmentation for the Spectral Images & Medical Images.