Search

With the development of infrared dual-band focal plane detector in the field of engineering application, the research on image fusion algorithm of dual-band infrared thermal imager based on engineering application becomes more and more important. Because the human eye can only distinguish dozens of gray scales, but can distinguish thousands of color scales. At the same time, gray images can only store one-dimensional data, while color images can contain three-dimensional data, which expands the amount of information by tens of thousands of times, and is easier to be applied to the processing and processing of machine vision related work in the field of detection, search, recognition and tracking. Therefore the research on color image fusion is becoming more and more important. At present, color image fusion is the main development trend of infrared dual-band image fusion. The mainstream color image fusion can be divided into four categories: color mapping, color transfer, color lookup table and neural network. The above four algorithms are difficult to take into account the color richness, environmental adaptability and human eye observation comfort. In this paper, the above problems, this paper proposes a class on medium wave infrared and long wave band color map image fusion algorithms, the algorithm is based on each source image pixel and image mean differences, the differences between the information more reasonably map to color print, the final fusion image has a better image perception and environment adaptability, and preserve more details. A typical dual-band infrared thermal imager application Scene was selected, and the subjective image comparison and objective index analysis were carried out by algorithm simulation and other mainstream color fusion algorithms, which proved that the algorithm was effective and feasible.

Summary Life scientists often desire to display the signal from two different molecular probes as a single colour image, so as to convey information about the probes’ relative concentrations as well as their spatial corelationship. Traditionally, such colour images are created through a merge display, where each greyscale signal is assigned to different channels of an RGB colour image. However, human perception of colour and greyscale intensity is not equivalent. Thus, a merged image display conveys to the typical viewer only a subset of the absolute and relative intensity information present in and between two greyscale images. The Commission Internationale de l'Eclairage L*a*b* colour space (CIELAB) has been designed to specify colours according to the perceptually defined quantities of hue (perceived colour) and luminosity (perceived brightness). Here, we use the CIELAB colour space to encode two dimensions of information about two greyscale images within these two perceptual dimensions of a single colour image. We term our method a Perceptually Uniform Projection display and show using biological image examples how these displays convey more information about two greyscale signals than comparable RGB colour space-based techniques.

This study presents an automatic, human perception based chroma-keying algorithm that extracts the objects of interest (i.e. foreground) from monochromatic background. Given an image to be chroma keyed, the global colour distribution and the local texture property are analysed in CIECAM02 colour appearance model. After the analysis, input image is automatically segmented into three parts: foreground, background, and uncertain regions. Afterwards, the background colour is propagated from known background to uncertain region by using interpolation functions; and the foreground colour is estimated based on global colour distribution and a linear cost criteria. The quantitative and perceptual comparisons on the matting results show that the proposed method can reliably remove the background region, correctly restore the intrinsic foreground colour, and accurately keep the fine details. In addition, the authors implement the proposed method on a heterogeneous parallel computing architecture which efficiently distributes the workload among different processors. The simulation results show that the foreground objects can be accurately extracted from high-definition and/or ultra-high-definition videos in real time.

This project includes the basic electronic concepts, programming language and compliers, and the use of microcontrollers. The main concepts of this project are the uses of microcontrollers and the programming language. In this project, paper colour detection system detects the different colours and display on LCD. C language is used to assign the function of internal processing of PIC. The source data table in the program had to be displayed the different colours displayed on the LCD screen.

Traffic light recognition is of great interest for advanced driver assistance systems and autonomous driving but still an unsolved problem. While a traffic light has few visual features for detection from camera images we believe the characteristic light represents a potentially very strong and stable feature. The traffic light is actively emitting light which is rarely influenced by weather or lighting condition. When using a color lookup table for an image segmentation-based object detector, the process of creating the lookup table is the crucial point. In this paper, we propose a method for generating a lookup table using real world data of a large dataset. The training data is sampled from labeled objects and stored as multisets. We contribute a frequency-based filtering method to clean the samples before using a k-nearest neighbor classifier to generalize. The result is stored as a three dimensional lookup table. The main contribution is a neighborhood-biasing technique that allows setting different operating points online without retraining. A challenging real world dataset containing several thousands of active lights is used to evaluate the process.

Separate colouring on a band of temperatures can improve the usage of thermogram. We propose multi-segments colour scale (MS scale) to provide the utility for separate colouring. MS scale uses handle and handle pair to divide itself into segments. Each segment then generates a colour sequence according to the colours selected at the handle/handle pair. With the use of colour handle and handle pair, MS scale can produce a variety of colour scales, and enable specific colouring on bands of temperatures. The matlab code for the colour scale and the corresponding demo video is located at http://jenh.co/2014/08/13/pseudocolours-for-thermography/ .

The head and neck region has a complex spatial and topological structure, three-dimensional (3D) computer model of the region can be used in anatomical education, radiotherapy planning and surgical training. However, most of the current models only consist of a few parts of the head and neck, and the 3D models are not detachable and composable. In this study, a high-resolution 3D detachable and composable model of the head and neck was constructed based on computed tomography (CT) serial images. First, fine CT serial images of the head and neck were obtained. Then, a color lookup table was created for 58 structures, which was used to create anatomical atlases of the head and neck. Then, surface and volume rendering methods were used to reconstruct 3D models of the head and neck. Smoothing and polygon reduction steps were added to improve 3D rendering effects. 3D computer models of the head and neck, including the sinus, pharynx, vasculature, nervous system, endocrine system and glands, muscles, bones and skin, were reconstructed. The models consisted of 58 anatomical detachable and composable structures and each structure can be displayed individually or together with other structures.

In this Letter, a near-field communication method using smart-phone liquid-crystal-display luminescence is proposed. The proposed system is composed of three parts: colour pattern generation, smartphone detection and colour pattern decoding. When the user inputs digital data such as text or numbers, the sender section converts the digital data into RGB colour pattern images using a previously prepared pattern generation method. The receiver section detects these colour pattern images using a camera, extracts the colour values and decodes them into original digital data. The data transfer accuracy to show that this system can work in real situations is tested.

We propose a novel method to fuse true colour images with monochromatic non-visible range images that seeks to encode important structural information from monochromatic images efficiently but also preserve the natural appearance of the available true chromacity information. We utilise the β colour opponency channel of the lαβ colour as the domain to fuse information from the monochromatic input into the colour input by the way of robust grayscale fusion. This is followed by an effective gradient structure visualisation step that enhances the visibility of monochromatic information in the final colour fused image. Images fused using this method preserve their natural appearance and chromacity better than conventional methods while at the same time clearly encode structural information from the monochormatic input. This is demonstrated on a number of well-known true colour fusion examples and confirmed by the results of subjective trials on the data from several colour fusion scenarios. Introduction The goal of image fusion can be broadly defined as: the representation of visual information contained in a number of input images into a single fused image without distortion or loss of information. In practice, however, a representation of all available information from multiple inputs in a single image is almost impossible and fusion is generally a data reduction task. One of the sensors usually provides a true colour image that by definition has all of its data dimensions already populated by the spatial and chromatic information. Fusing such images with information from monochromatic inputs in a conventional manner can severely affect natural appearance of the fused image. This is a difficult problem and partly the reason why colour fusion received only a fraction of the attention than better behaved grayscale fusion even long after colour sensors became widespread. Fusion method Humans tend to see colours as contrasts between opponent colours and an improvement in visibility of structures from the monochrome can be achieved when they are used to encode a single HVS colour dimension consistently. The lαβ colour system effectively decorrelates the colour opponency and intensity channels and manipulating one causes no visible changes in the others. Colour fusion can be achieved by fusing one of the colour opponency channels with the monochrome image. We use the Laplacian pyramid fusion known to be one of the most robust monochrome fusion methods available. The Laplacian, also known as the DOLP (difference of low-pass) pyramid is a reversible multiresolution representation that expresses the image through a series of sub-band images of decreasing resolution, increasing scale, whose coefficients broadly express fine detail contrast at that location and scale. A simple fusion strategy creates a new fused pyramid by copying the largest absolute input coefficient at each location. The β channel of the lαβ space represents the red-green opponency and we base our fusion on encoding this channel of the colour input with the monochrome image. This causes warmer objects (lighter in IR) to appear redder in the fused image. The fusion proceeds in several steps. Initially we transform the colour input RGB image into the lαβ image. Monochrome fusion is then performed by decomposing the β image and the normalised monochrome into their Laplacian pyramid representations. We use the select max strategy to construct the fused pyramid but we only apply this to a small number of higher resolution pyramid sub-bands. Larger scale features in lower resolution sub-band images that constitute the natural context of the scene are sourced entirely from the colour image (β). This ensures that well defined smaller objects from the IR image are transferred robustly into the fused image as well as the broad scene context from the colour input. Reconstructing the fused pyramid produces the fused β image which is combined with the original l and α channels of the colour input to produce the fused RGB colour image. Edge Emphasis We encode only the β channel which has only a fraction of the overall colour signal power (most is in the intensity channel) so the contrast of the monochrome image structures is still relatively modest in the fused image. We can improve their visualisation using a relatively simple effect of gradient outline enhancement. Initially, we extract gradient information from the monochrome image using 3x3 Sobel edge operators. The responses to horizontal and vertical Sobel templates, sx and sy, are combined to evaluate gradient magnitude at each location. To enhance the structure visualisation, prior to fusion, to the monochrome input we add its gradient magnitude image. The enhanced monochrome image is well behaved as the used gradient filters are linear, and is used directly as the input into monochrome fusion. The gradient magnitude image effectively captures the primal sketch of the scene and encoding an opponency channel with this information improves the visualisation of the structural outline of the monochrome input in the colour fused image. Results and Conclusion A new “β fusion” colour image fusion method is presented that successfully both visualises important structure information from the monochrome input and preserves the natural appearance of the true colour input. Colour fusion is performed in the lαβ colour space known to decorrelate main colour opponencies seen by the human visual system. We chose the β channel representing the red-green opponency of the true colour image to encode structural information from the monochrome input by fusing them using modified Laplacian pyramid fusion. The visualisation of important structures from the monochrome input can be improved through a simple structure encoding step using its gradient information. The method is naturally extended to video fusion. The proposed fusion methods produce colour fused images with significantly better visualisation of important information from the monochrome input while almost entirely preserving the natural appearance of the true colour input. This was demonstrated on a number of well-known colour fusion examples and measured using subjective trials on the data from multiple surveillance scenarios.

Color space based modulation (CSBM) schemes have the potential to replace conventional intensity based modulation in visible-light communication (VLC) systems. This study attempts to address three major issues of the CSBM scheme in VLC systems, which are (1) performance improvement of CSBM as compared to conventional intensity based modulation, (2) intensity control, and (3) adaption to color variation. We improve the bit error rate (BER) performance of the CSBM scheme by incorporating an error boundary decision method. We analyze our proposed scheme by varying color during communication and introduce a color lookup table in order to obtain a more accurate target color at the receiver. We evaluate the CSBM scheme when the transmitter sends and does not send the target color information, and we also consider the effect of channel coding on the system performance. The computer simulation results, assuming an additive white Gaussian noise (AWGN) channel, demonstrate an uninterrupted communication system under intensity and color variation.

Today image processing and utilizations based on RGB-D data have huge interest in various research areas. Regarding segmentation, colour correction, depth information, comprehensive description of 3-Dimension scenes in RGB Depth data provides useful hints from another dimension to segregate substance in close colours. Here, a novel method is discussed to resolve such mixed, lost, and noise level increase of pixel problems of the RGB-D system and has been simulated using MATLAB software. Apart of colour correction images is secured by using hybrid cryptography method. The advantage of MATLAB is, it is widely available, continuously updated and has wider reach. In our approach, mixed pixel areas are detected using super pixel segmentation of colour and depth, and consolidate them to lost pixel regions. In addition, a hybrid filter is used to remove noise level of pixels. Experimental results shows the proposed method gives better performance on Mean, Intensity, Area, Perimeter, security.

This paper presents a steganographic scheme based on the RGB colour cover image. The secret message bits are embedded into each colour pixel sequentially by the pixel-value differencing (PVD) technique. PVD basically works on two consecutive non-overlapping components; as a result, the straightforward conventional PVD technique is not applicable to embed the secret message bits into a colour pixel, since a colour pixel consists of three colour components, i.e. red, green and blue. Hence, in the proposed scheme, initially the three colour components are represented into two overlapping blocks like the combination of red and green colour components, while another one is the combination of green and blue colour components, respectively. Later, the PVD technique is employed on each block independently to embed the secret data. The two overlapping blocks are readjusted to attain the modified three colour components. The notion of overlapping blocks has improved the embedding capacity of the cover image. The scheme has been tested on a set of colour images and satisfactory results have been achieved in terms of embedding capacity and upholding the acceptable visual quality of the stego-image.

Today image processing and utilizations based on RGB-D data have huge interest in various research areas. Regarding segmentation, colour correction, depth information, comprehensive description of 3-Dimension scenes in RGB-Depth data provides useful hints from another dimension to segregate substance in close colours. Here, super pixel segmentation based a new method is discussed to resolve such mixed, lost, and noise level increase of pixel problems of the RGB-D system and has been simulated using MATLAB software. The advantage of MATLAB is, it is widely available, continuously updated and has wider reach. In our approach, mixed pixel areas are detected using super pixel segmentation of colour and depth, and consolidate them to lost pixel regions. The consolidated regions are replete with neighbouring depth information based on a Robust Edge-Stop (RES) Function; distance transform values of colour edge pixels are used to form this function. In addition, a hybrid filter is used to remove noise level of pixels. Experimental results shows the proposed method gives better performance on Mean, Intensity, Area, Perimeter, Centroid, and Diameter.

In the multi-color printing, the four inks used are cyan, magenta, yellow and black, which have different response in the near infrared spectral area, in which the black with CMK color inks have no response while the black with K color ink has a good response. But two kinds of black are both visible and equivalent under daylight. Based on this principle and gray component replacement (GCR) theory, this paper introduces a black component generation method for black ink hiding infrared security image. Firstly, we establish a color lookup table for GCR and find the substitution relation between K and CMY. Secondly, the source image is color-separated with the mask which is the security image. The source image is firstly converted to CMY color format data words. The black value of each pixel for security image is detected and treated as GCR value of the corresponding pixel for source image. Then CMY values are found in GCR LUT and subtracted from three primary CMY values. And an adjustment coefficient is employed for color correction. Thirdly, the color separation image is printed. In daylight the visual effect of printed image is the same as that of source image, but the security image is detected by IR camera. Finally, a series of experiments are performed on HP Color LaserJet CP2035 printer. Experimental results show that the proposed method is promising and feasible for black ink hiding IR security image.

This paper presents a simple and efficient index coding method to reduce the storage cost for indexed colour images. The method exploits the high degree of similarity among neighbouring indices by using a sorted palette. The indices that were compressed by the colour image quantisation are classified into three categories. Every category employs a different encoding rule to process the indices. To compress the indices of the second category, two optional approaches are provided. Experimental results show that the proposed method can significantly cut down the number of bit rates without incurring extra image degradation.

New generation vessels are equipped with fire detecting sensors; however, fire may not immediately be detected if it is far away from the sensors. The fire process therefore cannot be recorded. A video-based fire alarm system is developed to overcome the drawbacks of traditional fire detection equipment. This paper presents a video-based flame and smoke detection method for vessels. For flame detection, the dominant flame color lookup table (DFCLT) is created by using the fuzzy c-means clustering algorithm. The changed video frames are automatically selected and the changed regions deduced from these frames. An elementary, medium, or emergency flame alarm is then triggered by comparing the pixels of changed regions with the DFCLT. The changed video frames are automatically selected for smoke detection. The changed regions are deduced from these frames. If the shape of the changed region conforms to the characteristic which the top area is wider than the bottom area, a dangerous smoke alarm is sounded. The experimental results show that the proposed fire detection approach can detect dangerous flames and smoke, effectively and efficiently.

The colour assessment ability of the computer vision system is investigated and the data are compared with colour measurements taken by a conventional colorimeter. Linear and quadratic models are built to improve currently used methodology for the conversion of RGB colour units to L * a * b * colour space. For this purpose, two innovative ideas are proposed and tested. First, substantial amount of colour tones is generated to cover as many points in the colour space as possible. Secondly, the colour space is calibrated separately, whereas in previous research in the literature, the colour space is calibrated simultaneously. It is found that the RGB colour units to L * a * b * colour space transformation approach proposed in this study is more logical and more accurate, and the prediction performance of the quadratic model is superior over the linear model.

In this paper, we present a new multi-volume rendering technique (i.e., progressive fusion) to combine 3D anatomical structures from B-mode imaging and flow information from power Doppler imaging. A post-fusion technique, in which B-mode and power Doppler volumes are independently rendered and then fused based on alpha blending, is typically used in 3D power Doppler imaging. However, it has limitations in preserving the spatial relationship (i.e., depth order) between tissue structure and vasculature since they are rendered independently and then merged. With the proposed progressive fusion, B-mode and power Doppler volumes are composited together while rendering by sharing the opacity values. After compositing, two rendered frames are blended by utilizing a 2D color lookup table designed to fuse two properties (i.e., tissues and blood flows). We have evaluated the progressive-fusion multi-volume rendering method with the phantom and in vivo data acquired using a commercial ultrasound machine (EUB-8500, Hitachi Medical Corporation, Japan) with a 3.5 MHz mechanical probe. From the preliminary study, we have found that the new progressive-fusion method can better retain and display the spatial relationship between tissue structure, vasculature and their corresponding depth order.

This article proposes an original method for grading the colours between different images or shots. The first stage of the method is to find a one-to-one colour mapping that transfers the palette of an example target picture to the original picture. This is performed using an original and parameter free algorithm that is able to transform any N-dimensional probability density function into another one. The proposed algorithm is iterative, non-linear and has a low computational cost. Applying the colour mapping on the original picture allows reproducing the same ‘feel’ as the target picture, but can also increase the graininess of the original picture, especially if the colour dynamic of the two pictures is very different. The second stage of the method is to reduce this grain artefact through an efficient post-processing algorithm that intends to preserve the gradient field of the original picture.

Digital colour image enhancement using a homomorphic filter provides good dynamic range compression, but it fails in colour rendition. In this letter, we propose to perform natural colour rendition in a digital colour image that is enhanced by a homomorphic filter. A novel neural network learning algorithm, named Ratio rule, is used to carry out the natural colour rendition process. The Ratio rule learns to restore colour in an image by representing the colour relationships of each pixel in the original image as a line of attraction in the state space. The dynamics of the neural network is then used as an associative memory for recalling the natural colour characteristics of image pixels after homomorphic filtering. Several experiments on benchmark problems show that the new technique of image enhancement can generate natural colour in images from their original colour information.

Visual cryptography is different from traditional cryptography. That is, neither time-consuming computation nor complex cryptographic knowledge is needed. Stacking is the only operation required to recover a secret image, and the individual image does not give the hackers any information about the secret image. None of researches tried to deal with meaningful colour share transparencies. Hence, two methods are proposed for hiding a colour image in two meaningful colour share transparencies in this paper. To achieve this goal, the colour decomposition approach and halftone technology are first applied to cope with secret colour images. Then the concept of the human visual system is utilized to generate two colour meaningful sharing transparencies. To support various applications, two variants are presented. The first proposed method, method-1, is suitable for simple colour images, and the second, method-2, provides better visibility of complex colour images.

This paper describes the development of a non-contact system for measuring the colour of printed material at web speeds, in addition to gravure-printed dot feature recognition. The system proposed uses a non-contact spectrophotometer based on a holographic grating, in conjunction with a conventional monochrome area scan camera, from which colour spectral data are extracted, while a xenon flash is used to illuminate colour samples. Dot features are captured by a magnification lens, in conjunction with a progressive scan camera. Software and hardware details of the system are given, together with the underlying mathematics for colour space conversion and measurement. Conversion equations from X, Y, Z chromaticity coordinates to the RGB system are presented, and also equations to convert from the L* a* b* colour space to X, Y, Z chromaticity coordinates. Experimental results are presented whereby the non-contact spectral system is shown to perform to a colour tolerance exceeding that of conventional colour video systems, and where performance of the dot feature system is comparable with traditional static examination methods using a microscope.

Generally speaking, the human eye is less sensitive to changes in colour for varicoloured images. As for an image with high colour intensities, people find it difficult to perceive any change in colour if the colour intensities of some pixels are adjusted by a small quantity. Based on these two observations, this paper classifies the original image into blocks of three types, which are complex blocks, white blocks and smooth blocks. Thus, when watermarking takes place, the colours of the pixels are varied depending on the block type; as a result, someone can take good care of the quality of both the embedded image and the extracted watermark. The experimental results show that the proposed watermarking system has high robustness against a variety of aggressive image processing, including blurring, sharpening, cropping and JPEG lossy compression.

Mobile devices such as smartphones are going to play an important role in professionally image processing tasks. However, mobile systems were not designed for such applications, especially in terms of image processing requirements like stability and robustness. One major drawback is the automatic white balance, which comes with the devices. It is necessary for many applications, but of no use when applied to shiny surfaces. Such an issue appears when image acquisition takes place in differently coloured illuminations caused by different environments. This results in inhomogeneous appearances of the same subject. In our paper we show a new approach for handling the complex task of generating a low-noise and sharp image without spatial filtering. Our method is based on the fact that we analyze the spectral and saturation distribution of the channels. Furthermore, the RGB space is transformed into a more convenient space, a particular HSI space. We generate the greyscale image by a control procedure that takes into account the colour channels. This leads in an adaptive colour mixing model with reduced noise. The results of the optimized images are used to show how, e. g., image classification benefits from our colour adaptation approach.

Monochrome images are often converted to false-colour images, in which arbitrary colours are assigned to regions of the image to aid recognition of features within the image. Criteria for selection of colour palettes vary according to the application, but may include distinctiveness, extensibility, consistency, preference, meaningfulness and universality. A method for defining a palette from colours on the surface of a reference gamut is described, which ensures that all colours in the palette have the maximum chroma available for the given hue angle in the reference gamut. The palette can be re-targeted to a reproduction medium as needed using colour management, and this method ensures consistency between cross-media colour reproductions using the palette.

Tuning the colour thresholds within robot soccer is laborious, and sensitive to changes in illumination. An algorithm for automatic gain control and white balancing within the camera is described. This significantly reduces the effects of the variations in lighting on the thresholds. Next, a new colour space is proposed which maximises the hue separation of the different coloured regions to improve the colour discrimination. Finally several automatic segmentation techniques are briefly discussed. All of the algorithms are designed to operate on data directly streamed from the camera, enabling a low latency FPGA implementation.

Colour identification is a common task for many applications in computer vision. Many existing colour identification methods are application dependent. Therefore, they could only operate in certain environments and their performances are highly dependent on the lighting condition and quality of the images. In this paper, a novel colour identification method is introduced using an adaptive hue based colour model. The proposed method firstly analyses the image by collecting a variety of information from the given image. It is then followed by a colour space selection scheme in which the most suitable hue based colour space is selected based on properties obtained through image analysis. Finally, the image is filtered through the channels of the selected colour model to identify the target colour. Tests on various images from different datasets have shown that this method is capable of extracting the required colour effectively from images with varying qualities.

This paper proposes a simple and fast method for adapting colour lookup tables to lighting changes in real-time. The method adjusts the classified colour space regions keeping both their surface area and volume constant. Two variations of the method were compared and tested in a RoboCup soccer setting. Detection success rate was measured as a function of the speed and magnitude of hue change to the lighting environment. Compared to a static lookup table, these experimental results show improved robustness against lighting changes for detection of coloured objects.

With the advent of digital computers colour reproduction systems have been very quick to adopt digital image processing in every field of image reproduction. Digital information is based on digits representing data in physical quantities of binary digits of 0 and 1. In any colour image reproduction in any system the fundamental logic is based on balancing basic colours to produce ‘grey balance’, producing pleasant shades of basic colours called ‘tone reproduction’ and reproduce as clean colours as possible to avoid muddy colours in final reproduction is called ‘colour correction’.

Natural feeling color image fusion technology is an important direction in image processing. This paper first summarizes contributions from both domestic and foreign researchers. Then according to the theory of 2D color-lookup-table, it puts forwards the multiple reference images fusion algorithm, and gives the results of using two and three reference images. The comparing and analyzing of those results proves the validity of the algorithm.

For video compression, the RGB signals are usually converted at the source to a perceptual colour space, followed by chroma sub-sampling, for coding efficiency. This is based on the typically higher human visual system sensitivity to the luminance than chrominance of image signals. However, there are specific applications that demand carrying the full RGB signals through the transmission chain, which may also benefit from lossless colour transforms, for efficient coding. In either case, the best colour transform function is noted to be content dependent, although fixed transforms are typically adopted for convenience. This paper presents a method of dynamically adapting this colour transform function for each picture block, using a class of low complexity lifting based schemes. The performance of the proposed algorithm is compared with a number of fixed colour transform schemes and shows a significant compression gain over native RGB coding and YC o C g transform.

This paper describes the development of a non-contact system for measuring colour of printed material at web speeds. The system proposed uses a non-contact spectrophotometer based on a holographic grating, in conjunction with a conventional monochrome area scan camera, from which colour spectral data is extracted, whilst a xenon flash is used to illuminate colour samples. Software and hardware details of the system are given, along with the underlying mathematics for colour space conversion and measurement. Conversion equations from X, Y, Z chromaticity co-ordinates to the RGB system are presented, and also equations to convert from the L ∗ a ∗ b ∗ colour space to X, Y, Z chromaticity co-ordinates. Experimental results are presented whereby the non-contact spectral system is shown to perform to a colour tolerance exceeding that of conventional colour video systems.

Camera‐based colour measurement systems have application where it is difficult to apply a contact instrument, or where the contact instrument, because of its fixed lighting geometry, does not provide the necessary data. The paper describes the principles, procedures and precautions necessary for camera based colorimetry. Equations are presented to convert digitised R G and B values, which are system specific, into CIE X Y Z colour values, an international standard method of colour specification. It is reported that colours spaced equally in RGB values appear to have visual colour difference steps varying in size by as much as 30:1 through the different regions. The poorest colour resolution occurs for RGB values that represent dark or strongly coloured shades. At least ten bits per channel digitisation of RGB values is required to achieve a colour resolution similar to the human visual system for dark, intense shades.

Application of machine vision for automated visual inspection of cotton quality in textile industries is proposed. This automated visual inspection can be done by analysis of cotton image. The analysis involves color discrimination between cotton and impurities. Colors can be quantitatively described by color coordinates. Generally, human observers cannot discriminate between two colors if they can be described within an isodiscrimination contour. To inspect the quality of cotton, the cotton image will be analyzed for impurities using this isodiscrimination controu. The issues of signal processing and illuminations will also be discussed. These factors can contribute to errors in the inspection system. Also, a method for significantly improving speed of image processing will be mentioned.

Memory for colours presented in isolation was compared with that for colours presented as part of a clip-art image or as part of a non-meaningful, Mondrian-like image. The area under the Receiver Operating Characteristic for deciding that two colours, presented at different delay intervals, were the same or different provided an index of memory for colour. The provision of a context reduced the decay of memory, regardless of whether the context was meaningful (clip-art images) or non-meaningful (Mondrian-like images). The result was seen as a generalisation of the auditory phenomenon of profile analysis, in which memory for the amplitude of a single component of a complex sound is more stable than that for the component in isolation.

In this chapter we discuss how images are stored, starting off with the simplest image data, 1 bit per pixel. Grayscale, 8-bit images are considered next, and we consider the problem of how to actually send images to a printer and what problems and solutions this entails. RGB 24-bit color images are then explored, looking at the three color planes as images in their own right, and capturing RGB information content via a three-dimensional histogram of colors is introduced. Returning to 8 bits, the concept of a color palette is discussed; but this then brings up the problem of just what colors shall we distinguished by inclusion in the palette. To this end, the problem of how to devise a color lookup table is considered, with the Median-Cut algorithm brought forward as one solution. Popular file formats, GIF, JPEG, PNG, TIFF, PDF and so on are introduced. The GIF format is simple but representative, and for this reason it is set out in some detail. To wrap up the discussion the PTM (Polynomial Texture Mapping) technique and format is included, with the capability of virtually exploring the surfaces of cultural heritage precious objects.

Still, much of computer vision is predicated on greyscale imagery. There are good reasons for this. For much of the development of computer vision greyscale images were all that was available and so techniques were developed for that medium. Equally, if a problem can be solved in greyscale - and many can be - then the added complexity of starting with 3 image planes as oppose to 1 is not needed. But, truthfully, colour is not used ubiquitously as there are some important concepts that need to be understood if colour is to be used correctly. In this chapter I summarise the basic model of colour image formation which teaches that the colours recorded by a camera depend equally on the colour of the prevailing light and the colour of objects in the scene. Building on this, some of the fundamental ideas of colorimetry are discussed in the context of colour correction: the process whereby acquired camera RGBs are mapped to the actual RGBs used to drive a display. Then, we discuss how we can remove colour bias due to illumination. Two methods are presented: we can solve for the colour of the light (colour constancy) or remove it through algebraic manipulation (illuminant invariance). Either approach is necessary if colour is to be used as a descriptor for problems such as recognition and tracking. The chapter also touches on aspects of human perception.

This research presents a novel colour management system within an Image Signal Processor pipeline that encompasses the colour constancy problem and reproduces perceptually pleasing colours in the final image produced by an ISP. The proposed colour management system includes a new machine learning based automatic white balancing algorithm, an intelligent colour matrix balancing method for achieving colour constancy in non-orthogonal camera colour functions and a cooling-warming approach based on Kruithof curve to achieve pleasing colours in the final image produced by the ISP.

A look-up table algorithm is described for enhancing the luminance resolution at the expense of the colour resolution in digital displays of colour images. For colour displays with a look-up table resolution of 8 bits/gun, the algorithm provides a luminance resolution of 11-12 bits without loss of spatial or temporal resolution. This improvement can reduce the minimum luminance step in the mid-luminance range from 1.5% to > 0.2% with no additional hardware.

That image colours depend equally on the colour of surfaces and the colour of the prevailing light means that raw RGB pixel values are not correlates of object features and so, cannot be used directly in applications such as recognition and tracking. By estimating the colour of the light and then removing the colour bias due to illumination, image colours can be made to correlate with object reflectance. Yet, even the best algorithms can fail badly. The problem is made even harder in the many typical scenes where there are multiple lights, e.g. sun (yellow) and shadow (blue). We demonstrate that the second, harder case actually provides the basis of a novel solution strategy: an illumination edge in an image provides a powerful cue to determining the underlying material surface. This paper makes three important technical contributions. First, we take an existing analytic one-surface two-lights estimation algorithm (which assumes that lights lie on a line in chromaticity space) and re-engineer it as a voting algorithm which relaxes the prior assumption about where lights might lie. Second, we model illumination change using a 3 × 3 linear matrix as opposed to the more approximate diagonal model that is used in the prior art for this problem. Being able to more accurately account for illumination change provides a step change in the performance of our method. Lastly, we re-engineer a prior art shadow-edge generator to suit this problem.

The use of machine vision technology is being investigated at VTT for improving the colour quality and productivity of web offset printing. The visual inspection of colour quality is performed by a colour CCD camera which traverses the moving web under a stroboscopic light. The measuring locations and goal values for the colour register, the ink density and the grey balance are automatically determined from the PostScript™ description of the digital page. A set of criteria is used to find the most suitable spots for the measurements. In addition to providing data for on-line control, the page analysis estimates the zone wise link consumption of the printing plates as a basis for presetting the ink feed. Target calorimetric CIE-values for grey balance and critical colours are determined from the image originals. The on-line measurement results and their derivations from the target values are displayed in an integrated manner. The paper gives test results of computation times, measurements of register error with and without test targets and the colour measuring capabilities of the system. The results show that machine vision can be used for on-line inspection of colour print quality. This makes it possible to upgrade older printing presses to produce a colour quality that is competitive with more modern presses.

Automatic inspection systems based on image processing have been used in manufacturing systems. However, in the case of systems using colour images, colour signals fed into the system through a camera depend upon the inspected objects and on the light sources. Accordingly, it is necessary to convert the colour signals received under the illumination at the time of inspection into those that would be received when using a standard light source to achieve a reliable and versatile inspection system that can operate stably under a wide variety of light sources. In this paper, a colour image input system with white balancing using a neural network is proposed, and compared with that used in ordinary video cameras by means of computer simulation. This system is applicable as the pre-processor of colour signals for automatic inspection systems based on colour image processing.

Colour is an important aspect of art. Not only does it give richness\ud to images, but it always provides a means to highlight certain objects.\ud This idea of spot colour has been used extensively in both fine\ud art and commercial illustrations. Many non-photorealistic rendering\ud (NPR) algorithms produce grayscale or monochromatic images\ud with low saturations. In this paper we introduce the idea of spot\ud colour to NPR and propose a simple and automatic algorithm to add\ud spot colour to these rendering styles. The hue is thresholded into\ud colour layers and the most appropriate layer is automatically determined\ud based on factors such as layer region shape and salience. We\ud also consider using an edge-based criterion to colourise the background,\ud which is an effective means of making the foreground stand\ud out. We demonstrate the effectiveness of our approach by adding\ud spot colour to a diverse set of NPR styles.

We describe a system for storing, transmitting, and presenting digital movies in a computer network. The hardware used in the system is standard hardware, as found in typical workstations today; no special hardware is required, but if available it can be used to provide better performance. The XMovie system has several innovative features. First, it contains a new algorithm for the gradual adaptation of the color lookup table during the presentation of the movie to ensure optimal color quality on low-end workstations. Second, it is a multistandard system supporting the compression techniques MPEG, Motion JPEG, and a newly developed extension to the well-known Color Cell Compression method. Third, it contains AdFEC, a new adaptable forward error correction method for our movie transmission protocol.

A new algorithm using invariant spectral features for segmenting colour images is presented in this paper. Input data are three primary images obtained from a colour sensor. The input colour image is transformed to IHS (Intensity, Hue, Saturation) colour space. This colour space transform compensates for illumination variations and delivers image pixel values with low variance for individual colour regions, hence contributing to simplified segmentation. The hue and saturation images are then separately filtered and combined. The resulting image is segmented by means of a threshold process. An opening operation on the segmented image completes the algorithm. Experimental results obtained for several images are presented. Issues related to illumination and sensors are also addressed.

This paper presents the design and VLSI implementation, on a single chip, of a new real-time colour space converter, which performs the transformation of the RGB colour coordinates to the intensity, hue and saturation (IHS) colour space. This high speed VLSI chip is designed to work at 13.3 MHz and can process high resolution colour images of up to 697 pixels per horizontal line. Its minimum rate of operation is 133 MIPS. The high speed of operation was achieved by pipelining the data in a vector fashion. The chip was implemented using a DLM, 1.0 μm, N-well, CMOS process provided by European Silicon Structures and the die size dimensions for the chip are 5.669 mm x 6.144 mm = 34.830 mm2. The chip is intended to be used as the front end of colour machine vision systems in inspection tasks and in autonomous applications, where the need for short processing times is crucial. Finally, simulation results exhibiting the usefulness of the IHS colour space transform using the proposed ASIC and real colour image data are also presented.

This paper presents a technique for the recognition of objects which are both complex and coloured. The transformation of the colour cluster is designed to obtain a colour set which makes it possible to use the colour cluster information for n- tuple classification. Generally, if two objects have different colour patterns, their colour clusters will be different. Therefore, the colour cluster is an important characteristics for coloured object recognition. Further processing takes place, where some colour planes are sliced from the colour set, and every plane is sampled separately and applied to a classifier based on the n- tuple technique of pattern recognition. The colour cluster called ‘pure colour information’ is position invariant. Therefore, the system has one great advantage of position invariant inspection for coloured objects. The system interrogation time is always the same for whatever the complexity of the object, which is another advantage of the technique for the inspection of complex objects. Preliminary results using these techniques have shown promise in the classification and inspection of various artefacts.

In an object-based analysis-synthesis coder each image of an image sequence is subdivided into arbitrarily shaped moving objects. The objects are described by three parameter sets defining their motion, shape and colour, where the colour parameters denote the luminance and chrominance values of the object surface. In order to increase the efficiency of the coder, a predictive technique for the coding of colour parameters is presented using image pattern prototypes for prediction. The prototypes are extracted from previously coded images and identically stored at the coder and the decoder. They can be used for the prediction of the colour parameters of an object by applying a suitably defined mapping function. The prediction error is coded by an adaptive combination of DCT and intraframe DPCM. For the reconstruction of the colour parameters at the decoder the address of the image pattern prototype and the parameters of the mapping function have to be transmitted as side information. Prototype prediction is only applied, when its data rate is lower than that of an extended hybrid scheme. At a transmission data rate of 61 kbit/s prototype prediction increases the PSNR by 2 dB in the areas of colour update. >

This paper presents a new copyright protection technique using colour watermarks. The watermark is a RGB colour image where each pixel is represented by 24 bits. The RGB colour watermark is first preprocessed by converting it into binary sequence. The preprocessed colour watermark (binary image) is embedded multi-times into the green channel of the RGB host image by modifying the very low frequency coefficients of the DCT transformation. The new technique can resist classical attacks such as JPEG compression, low pass filtering, median filtering, cropping, and scaling attacks. Moreover, the recovery method is blind since it does not need the original host image for extraction. The proposed method implemented using a graphical user interface (GUI) to open, select, attack, evaluate and display the results of both the host image and colour watermark image. The implemented GUI is much easier for the end users to learn more than the traditional algorithms that need commands to be know and memorized.