Color Image Processing

Real-time image classification is one of the most challenging issues in understanding images and computer vision domain. Deep learning methods, especially Convolutional Neural Network (CNN), has increased and improved the performance of image processing and understanding. The performance of real-time image classification based on deep learning achieves good results because the training style, and features that are used and extracted from the input image. This work proposes an interesting model for real-time image classification architecture based on deep learning with fully connected layers to extract proper features. The classification is based on the hybrid GoogleNet pre-trained model. The datasets that are used in this work are 15 scene and UC Merced Land-Use datasets, used to test the proposed model. The proposed model achieved 92.4 and 98.8 as a higher accuracy.

Illumination invariance remains the most researched, yet the most challenging aspect of automatic face recognition. In this paper we investigate the discriminative power of colour-based invariants in the presence of large illumination changes between training and test data, when appearance changes due to cast shadows and non-Lambertian effects are significant. Specifically, there are three main contributions: (i) we employ a more sophisticated photometric model of the camera and show how its parameters can be estimated, (ii) we derive several novel colour-based face invariants, and (iii) on a large database of video sequences we examine and evaluate the largest number of colour-based representations in the literature. Our results suggest that colour invariants do have a substantial discriminative power which may increase the robustness and accuracy of recognition from low resolution images.

In this keynote the focus is on mapping out the landscape of benefits provided by multi-primary printing and on highlighting some of the major challenges that stand in the way of harnessing its potential. A brief introduction to three and four primary printing and the business needs for multi-primary printing is followed by an enumeration of potential benefits that printing with more than three primaries has. Next some of the key challenges of using larger numbers of primaries are discussed and finally examples are given of future opportunities opened by multi-primary printing.

Color  Image  quantization  (CQ)  is  an  important problem  in  computer  graphics,  image  and  processing. The  aim  of
quantization is to reduce colors in an image with minimum distortion. Clustering  is  a  widely  used  technique  for  color  quantization;  all colors  in  an  image  are  grouped  to  small  clusters.  In  this  paper,  we proposed a new hybrid approach for color quantization using firefly algorithm (FA) and K-means algorithm. Firefly algorithm is a swarm based algorithm that can be used for solving optimization problems. The  proposed  method  can  overcome  the  drawbacks  of  both algorithms  such  as  the  local  optima  converge  problem  in  K-means and  the  early  converge  of  firefly  algorithm.  Experiments  on  three commonly  used  images  and  the  comparison  results  shows  that  the
proposed  algorithm  surpasses  both  the  base-line  technique  k-means clustering and original firefly algorithm

Several efficient component and vector filters applicable for processing multichannel and color images are compared quantitatively. The cases of Gaussian and mixed Gaussian and impulsive noise are considered. It is shown that in situations typical for practice the best results are provided by a vector doublewindow modified trimmed mean filter and by a modified sigma filter applied to each component.

ABSTRACT Purpose: Evaluation of computer based color deficiency test. Materials and Methods: Two hundred and sixty seven volunteers have been checked using both traditional Ishihara plates and a computer diagnosis program using LCD monitors. Results: The prevalence of red green color vision deficiency (RG-CVD) was 8.75% of male participants, no female participants were diagnosed, both in the paper based test, and in the computer based test. Computer based test gave 100% sensitivity and 98.78% specificity. Conclusion: Presenting the computer based color deficiency test software on LCD screen can be used for screening of color vision deficiency with nearly similar sensitivity and specificity to the Ishihara test with the advantage reducing the cost through decreasing required resources over time,and decreasing the time to analyze the results.

The goal of this work is to propose a set of techniques based on Mathematical Morphology to simplify or segment color images. Such simplification is done spatially or by quantization. The proposed methods apply the connected operators and morphological segmentation theories or extend them to color images.

Image processing is an engineering field where stored image data is readily available for parallel processing. Basically data processing algorithms developed in sequential approach are not capable of harnessing the computing power of individual cores present in a single-chip multicore processor. To utilize the multicore processor efficiently on windows platform for color image processing applications, a lock-free multithreading approach was developed using Visual C++ with Microsoft Foundation Class (MFC) support. This approach distributes the image data processing task on multicore Central Processing Unit (CPU) without using parallel programming framework like Open Multi-Processing (OpenMP) and reduces the algorithm execution time. In image processing, each pixel is processed using same set of high-level instruction which is time consuming. Therefore to increase the processing speed of the algorithm in a multicore CPU, the entire image data is partitioned into equal blocks and copy of the algorithm is applied on each block using separate worker thread. In this paper, multithreaded color image processing algorithms namely contrast enhancement using fuzzy technique and edge detection were implemented. Both the algorithms were tested on an Intel Core i5 Quad-core processor for ten different images of varying pixel size and their performance results are presented. A maximum of 71% computing performance improvement and speedup of about 3.4 times over sequential approach was obtained for large-size images using four thread model.

Tomatoes are best known grown fruits in greenhouses. In order to decrease the high labor cost and increase the efficiency, recently many researches have tried to harvest tomatoes automatically. To help the harvesting robot picking the fruit precisely, the main challenges are to identify and localize red mature tomato fruit accurately. In this paper, we presented a robust and effective method for challenging task of segmentation and localization of on-plant tomatoes in the greenhouse, for automatic tomato harvesting purpose. A low-cost RGB-D Kinect V2 camera was used to accomplish the task. Color images and depth images of tomatoes on-plants in the greenhouse were captured in day time using Kinect V2 camera. Along with color and depth images, the corresponding color coordinates, depth coordinates, and camera coordinates were also saved. The proposed method was divided into two stages: fruit segmentation and fruit localization. At the first stage, an improved HSV algorithm and watershed algorithm were used to segment mature fruits from a complex background. An improved HSV algorithm was used to enhance the feature of red mature fruits. After that, Otsu method and morphological operations were used to segment mature fruits from the background and then watershed was implemented to further separate touched fruits. Subsequently, each detected fruit was separately labeled to calculate the pixel coordinates values in color image. At the second stage, a fusion of color and depth information was carried out on the basis of the mapping relationship between color, depth and camera coordinates, to map the obtained fruit pixel coordinates in color image with acquired 3D point cloud data. Then the corresponding point cloud data of each labeled fruit were generated. To remove the noise in the cloud data, a filtration process was performed, and the final point cloud was generated. To verify the proposed method, 18 pairs of images were experimented containing the different numbers of measured fruits with measuring distance, ranging from 0.58 to 1.50 m. The average fruit localization error rate was 1.2 cm for depth distance and 1.9 cm for ground distance. Final results showed that the proposed method has good performance and robustness for segmentation and localization and have good potential to be implemented on the tomato harvesting robot in the greenhouse.

In this paper we present an approach to convert
a gray scale image to colour using artificial neural network.
The problem of converting a gray scale to a colour is an
undetermined problem and has no unique solution. We
designed a neural network (feed forward) which is being
trained by sets of sample images using back-propagation to
recognize the pattern in RGB values and gray scale values.
This trained neural network is then used to convert gray scale
image to colour, reducing the computation time to less than a
second to 5-6 seconds according to the size of the image
which is much reduced compared to the other methods
described. Mean Square Error (MSE) was used to measure
the results quantitatively.

Conversion of a color image into a grayscale image
inclusive of salient features is a complicated process. The
converted grayscale image may lose contrasts, sharpness,
shadow, and structure of the color image. To preserve
contrasts, sharpness, shadow, and structure of the color image
a new algorithm has proposed. To convert the color image into
grayscale image the new algorithm performs RGB
approximation, reduction, and addition of chrominance and
luminance. The grayscale images generated using the
algorithm in the experiment confirms that the algorithm has
preserved the salient features of the color image such as
contrasts, sharpness, shadow, and image structure.

Identification of plant disease is very important to prevent the loss and keep the harvest healthy. Determination of plant disease via visual monitoring is difficult and time consuming. In this paper, we described a method of detection and classification of plant disease using image processing and machine learning techniques. We used standard images of leaves of several types of plants to test our method. Initially, our method segments the input image to isolate disease parts of the leaf. Then we obtain various features from the diseased affected segmented image. Finally, we classify leaves into healthy and disease types based on its features using Multiclass Support Vector Machine (SVM) classifier. Experimental results indicate that our method yields very high accuracy rate for detection and classification of plant diseases.

D igital still color cameras (DSCs) have gained significant popularity in recent years, with projected sales in the order of 44 million units by the year 2005. Such an explosive demand calls for an understanding of the processing involved and the implementation issues, bearing in mind the otherwise difficult problems these cameras solve. This article presents an overview of the image processing pipeline, first from a signal processing perspective and later from an implementation perspective, along with the tradeoffs involved.

Color image segmentation is a very emerging research topic in the area of color image analysis and pattern recognition. Many state-of-the-art algorithms have been developed for this purpose. But, often the segmentation results of these algorithms seem to be suffering from miss-classifications and over-segmentation. The reasons behind these are the degradation of image quality during the acquisition, transmission and color space conversion. So, here arises the need of an efficient image enhancement technique which can remove the redundant pixels or noises from the color image before proceeding for final segmentation. In this paper, an effort has been made to study and analyze different image enhancement techniques and thereby finding out the better one for color image segmentation. Also, this comparative study is done on two well-known color spaces HSV and LAB separately to find out which color space supports segmentation task more efficiently with respect to those enhancement techniques.

Prediction of lung cancer is most challenging problem due to structure of cancer cell, where most of the cells are overlapped each other. The image processing techniques are mostly used for prediction of lung cancer and also for early detection and treatment to prevent the lung cancer. To predict the lung cancer various features are extracted from the images therefore, pattern recognition based approaches are useful to predict the lung cancer. Here, a comprehensive review for the prediction of lung cancer by previous researcher using image processing techniques is presented. The summary for the prediction of lung cancer by previous researcher using image processing techniques is also presented.

In this paper, we propose a color image quantization algorithm based upon TBM. In this context, we consider that the color quantization problem can be viewed as clustering problem of the color-space into P clusters. Using TBM, we define a top-down evidential clustering algorithm which iteratively decreases the number of clusters of the color space into P clusters. This convergence is ensured using a novel criterion based upon the pignistic probability function. The P clusters provide the new reduced color palette and a quantized color image is computed. This quantization method is completely automatic and preserves the final result from any initial condition. Experiments on various images show the algorithm efficiency for color quantization and highlight the efficiency of TBM for color image processing.

Content based image retrieval (CBIR) is the task of searching digital images from a large database based on the extraction of features, such as color, texture and shape of the image. Most of the research in CBIR has been carried out with complete queries which were present in the database. This paper investigates utility of CBIR techniques for retrieval of incomplete and distorted queries. Studies were made in two categories of the query: first is complete and second is incomplete. The query image is considered to be distorted or incomplete image if it has some missing information, some undesirable objects, blurring, noise due to disturbance at the time of image acquisition etc. Color (hue, saturation and value (HSV) color space model) and shape (moment invariants and Fourier descriptor) features are used to represent the image. The algorithm was tested on database consisting of 1875 images. The results show that retrieval accuracy of incomplete queries is highly increased by fusing color and shape features giving precision of 79.87%. MATLAB ® 7.01 and its image processing toolbox have been used to implement the algorithm.

Image enhancement is a processing on an image in order to make it more appropriate for certain applications. It is used to improve the visual effects and the clarity of image or to make the original image more conducive for computer to process. Contrast enhancement changing the pixels intensity of the input image to utilize maximum possible bins. We need to study and review the different image contrast enhancement techniques because contrast losses the brightness in enhancement of image. By considering this fact, the mixture of global and local contrast enhancement techniques may enhance the contrast of image with preserving its brightness. There are many image contrast enhancement techniques such as HE, BBHE, DSIHE, MMBEBHE, RMSHE, MHE. BPDHE, RSWHE, GHE, LHE and LGCS. This paper focuses on the comparative study of contrast enhancement techniques with special reference to local and global enhancement techniques. Also proposed solution is identified to apply to this enhancement technique. This novel method will use in many fields, such as medical image analysis, remote sensing, HDTV, hyper spectral image processing, industrial X-ray image processing, microscopic imaging etc. than other techniques because these techniques improve the visual effects and clarity of the image with preserving its brightness. This is proposed system, implementation work is going on. The major goal of image contrast enhancement is to produce images without severe side effects at the same time maintain input mean brightness.

The essay deals with the problem of working women and hand colouring in Italy between the end of the 19th and the beginning of the 20th century. It moves from some general issues, such as the relation between the colouring trade and the other professions involved with the diffusion of modernity; the continuity with other colouring praxis (i.e. photography); the possibility of framing the phenomenon in a tradition of hidden working women. After a synthetic reconstruction of the general context, the text investigates working women and division of labour in the Roman laboratories of Cines between 1905 and 1910. The main sources of the investigations are journals, magazines and handbooks of photography and cinematography, together with other sources testifying the state of feminine work in Italy during this period.

Identification of plant disease is very important to prevent the loss and keep the harvest healthy. Determination of plant disease via visual monitoring is difficult and time consuming. In this paper, we described a method of detection and classification of plant disease using image processing and machine learning techniques. We used standard images of leaves of several types of plants to test our method. Initially, our method segments the input image to isolate disease parts of the leaf. Then we obtain various features from the diseased affected segmented image. Finally, we classify leaves into healthy and disease types based on its features using Multiclass Support Vector Machine (SVM) classifier. Experimental results indicate that our method yields very high accuracy rate for detection and classification of plant diseases.

The essay deals with the problem of women workers employed as hand colorists in Italy between the end of the nineteenth and the beginning of the twentieth century. It explores general issues such as the relation between the modern development of industrial coloring processes and the professions involved, the continuity between the coloring techniques in the film industry and those employed in other production fields (e.g. photography). The text subsequently investigates the situation of women workers and the gendered forms of labor division in the Roman laboratories of Cines between 1905 and 1910.

Colour correction has introduced a new idea of colour image in the contemporary mediascape. What kind of relation is established, therefore, between the visible colour and the one reproduced by digital images? In order to investigate the mimetic status of digital colours, this essay provides a preliminary survey on colour correction practices adopted by some recent films. | | | Nello scenario mediale contemporaneo, le possibilità manipolatorie introdotte dalla color correction pongono in modo nuovo il rapporto tra dimensione produttiva e riproduttiva dell’immagine a colori. Che tipo di rapporto si instaura dunque tra il colore visibile e il colore riprodotto dall'immagine digitale? Per indagare lo statuto mimetico del colore digitale, il saggio propone una riflessione preliminare sulle pratiche di color correction adottate in alcuni film recenti.

In this study, a new classification algorithm based on multilevel thresholding for color images has been proposed.
Initially, thresholds for each channel of color images are determined by using histograms and bee algorithm. Then,
the threshold values obtained are used for partition of RGB color space. Thus, pixels located in the relevant sub
cubes were assigned in the same cluster and subsequently the results have been obtained.

Several technologies have been developed in recent years to advance the state of the art of IR sensor systems including dual color affordable focal planes, on-focal plane array biologically inspired image and signal processing techniques and spectral sensing techniques. Pacific Advanced Technology (PAT) and the Air Force Research Lab Munitions Directorate have developed a system which incorporates the best of these capabilities into a single device. The 'NeuroSeek' device integrates these technologies into an IR focal plane array (FPA) which combines multicolor Midwave IR/Longwave IR radiometric response with on-focal plane 'smart' neuromorphic analog image processing. The readout and processing integrated circuit very large scale integration chip which was developed under this effort will be hybridized to a dual color detector array to produce the NeuroSeek FPA, which will have the capability to fuse multiple pixel-based sensor inputs directly on the focal plane. Great advantages are afforded by application of massively parallel processing algorithms to image data in the analog domain; the high speed and low power consumption of this device mimic operations performed in the human retina.

We present a method for segmentation of cells from color images of blood smears in the frame work of statistical modeling. The 2-part approach results in locating each of the white blood cells (WBC) and identifying the regions corresponding to the nucleus and the cytoplasm, in the given blood smear. The given RGB image is first converted to its Hue (H), Saturation (S), Value (V) equivalent. Each pixel is treated as a vector of the three dimensions namely H, S and V. The components are weighted to give more importance to the most distinguishing features. We segment by modeling each of the above mentioned regions by a distinct 3-D Gaussian distribution. In the first step, K-means
clustering is performed on the 3-D feature vectors. This results in partitioning of the image into distinct regions. The centroids and the variances obtained in the K-means step are used to initialize Gaussian parameters for Expectation-Maximization (EM) algorithm. The EM algorithm iterates between segmentation and parameter estimation till convergence. A total of 115 images of smears were analyzed using our algorithm and successful segmentation was achieved in 80% of the cells contained in the images. The most important feature of this technique is that there are no parameters to be tuned by the user.

This paper presents 2-(hydroxyl) quinoline lithium (Liq) used as an n-type dopant to improve white hybrid organic light-emitting diode (WHOLEDs) performance. The Liq doped tris(8-hydroxyquinolinato) aluminum (Alq 3) layer possessed enhanced electron injection, efficient hole and electron balance in the emitting layer, as one of the most essential issues for device applications. This work investigates the optimum recipe (Liq concentration and thickness) of Alq 3 :Liq n-type doped electron injection layer (EIL) for WHOLED devices by comparing the current density and efficiency results with conventional Alq 3 /LiF technique. A blocking layer or interlayer is inserted between emitting layer and EIL to avoid excitons quenched. In this work suitable material and optimum thickness for blocking layer are studied, a white small-molecular organic light-emitting diode (SM-OLEDs) based on a 1,3,5-tris (N-phenylbenzimidazol-2-yl) benzene (TPBi) stamping transfer process is investigated. The proposed stamping transfer process can avoid the complexity of the vacuum deposition process.

Watermarking on Digital Images have different methods, some of them deals with frequency domain. The Watermarking techniques using frequency domain produce better results either concerns with image quality or invisibility of the watermark. The digital watermarking is such that it provides copyright protection by information hiding. The researchers all over the world are still trying to find out the way to create robust Digital Image Watermarking Techniques. This research paper produces the new concept of efficient watermarking technique for the Color Image by combining both DWT (Discrete Wavelet Transform) and DCT (Discrete Cosine Transform). The result includes combine features of both transformations, so the Watermarking is more Robust than earlier approaches. The Color Image Watermarking is done by selecting one color component from RGB (Red, Green, and Blue) Components of Color Image. The Watermark embeds into any selected component and then again merges with other components. The R (Red) Component plays very important role to the present the color object as well as it is robust to the preserve information associated with it. So, in this research, R (Red) Component is selected to hide Watermark. This approach is more secure because the embedded watermark can only be extract from the Red Component after decomposing Watermarked Image into RGB components, other components (i.e. Blue and Green) does not contain any information about the Watermark.

Color constancy is one of the important research areas with a wide range of applications in the fields of color image processing and computer vision. One such application is video tracking. Color is used as one of the salient features and its robustness to illumination variation is essential to the adaptability of video tracking algorithms. Color constancy can be applied to discount the influence of changing illuminations. In this paper, we present a review of established color constancy approaches. We also investigate whether these approaches in their present form of implementation can be applied to the video tracking problem. The approaches are grouped into two categories, namely, Pre-Calibrated and Data-driven approaches. The paper also talks about the ill-posedness of the color constancy problem, implementation assumptions of color constancy approaches, and problem statement for tracking. Publications on video tracking algorithms involving color correction or color compensation techniques are not included in this review.

A novel quaternion color representation tool is proposed to the images and videos efficiently. In this work, we consider a full model for representation and processing color images in the quaternion algebra. Color images are presented in the threefold complex plane where each color component is described by a complex image. Our preliminary experimental results show significant performance improvements of the proposed approach over other well-known color image processing techniques. Moreover, we have shown how a particular image enhancement of the framework leads to excellent color enhancement (better than other algorithms tested). In the framework of the proposed model, many other color processing algorithms, including filtration and restoration, can be expressed.

Color information of natural images can be considered as a highly correlated vector space. Many different color spaces have been proposed in the literature with different motivations toward modeling and analysis of this stochastic field. Recently, color transfer among different images has been under investigation. Color transferring consists of two major categories: colorizing grayscale images and recoloring colored images. The literature contains a few color transfer methods that rely on some standard color spaces. In this paper, taking advantages of the principal component analysis (PCA), we propose a unifying framework for both mentioned problems. The experimental results show the efficiency of the proposed method. The performance comparison of the proposed method is also given.

This paper proposes COCOCLUST, a contour-based color clustering method which robustly segments and binarizes colored text from complex images. Rather than operating on the entire image, a ‘small’ representative set of color pixels is first identified using the contour information. The method involves the following steps: (i) identification of prototype colors (ii) a one-pass algorithm to identify color clusters that serve as seeds for the refining step using k-means clustering (iii) assignment of pixels in the original image to the nearest color cluster (iv) identification of potential candidate text regions in individual color layer and (v) adaptive binarization. We propose a robust binarization technique to threshold the identified text regions, taking into account the presence of inverse texts, such that the output image always has black text on a white background. Experiments on several complex images having large variations in font, size, color, orientation and script illustrate the robustness of the method.

On September 18th, the day before the IX Italian Colour Conference, the University of Florence (in the ex-church of S. Verdiana) is going to host CHROMA | Workshop on colour image between motion pictures and media, organized in cooperation with GdC – Associazione Italiana del Colore, Fondazione Sistema Toscana - Mediateca Regionale, Università di Firenze. CHROMA intent is to offer scholars and professionals, sharing the same interest towards colour in motion pictures and media, a framework where dialogue and discussion may start. At the end of the day, at Teatro del Sale owned by the notorious chef Fabio Picchi, dinner and a screening of Our colour’s memory (in cooperation with Home Movies) will take place; the latter is a review edit of colour home movies among 1930s and 1960s with live music. The event will end with a screening of Le voyage dans la lune by Georges Méliès (restored version by Lobster Films, Groupama Gan Foundation for Cinema and Technicolor Foundation for Cinema Heritage). For more information please refer to: http://chroma.di.unimi.it/

Info and registration: chroma@outlook.it

While there may be no point in arguing about taste, creative professionals make a living from sharing theirs. Making specific, individual color preferences that a creative professional knows how to achieve when creating content on a display also propagate into print is a significant challenge since it lacks real-time feedback. The present paper introduces a method for allowing creative professionals to use the tools they know and love to also personalize the color behavior of their devices. This is achieved by analyzing color changes applied to images and applying them to a device's ICC profile. As a result the personalized device results in customized color behavior regardless of the workflow used. The paper describes the ICC profile transformation algorithm in detail and provides a color error analysis of its performance.

This paper presents two different approaches in color image decomposition domain with Bidimensional Empirical Mode Decomposition (BEMD). The first approach applies the BEMD on each channel separately and the second is based on interpolation of each channel in the sifting process. The comparison of two approaches shows the same performance of each approach in terms of visual quality, but they do not provide the same results in execution time which presents the most important criteron in real time applications. It was shown that the second BEMD approach based on interpolation of each channel in the sifting process, gives a gain in the point of view the execution time.