Search

In general, as people age the functions of their sensory organs deteriorate. For example, in the eye of an elderly person, the visual field becomes dark owing to a decrease in the amount of light reaching the retina. Moreover, objects appear blurred due to a decrease in contrast sensitivity. In particular, the contrast sensitivity decreases remarkably as the spatial frequency increases. In this study, we propose an image enhancement method that considers the visual characteristics of elderly persons. The proposed method combines lightness correction and amplification of high spatial frequency components to obtain images with optimized visibility for elderly persons. In particular, lightness enhancement that guarantees color gamut via tone mapping is conducted. In addition, image sharpening is applied based on the inverse characteristic function of the contrast sensitivity curve. The experiments were performed using several digital color images to confirm the effectiveness of the proposed method.

Images of the same object taken by multiple different cameras should have the same color reproduction. However, the images sometimes show different color reproduction due to the individual differences of cameras or internal camera parameters automatically determined when the images are taken. Conventional color transfer methods can be used for unifying the color reproduction of images by transforming the color distribution of an image to that of a reference image. However, conventional methods do not always lead to a good color reproduction and sometimes result in the loss of color impression of original images. In the present paper, we propose a color calibration method for images of the same object taken by different cameras. Two color transfer methods are combined to realize color calibration without the loss of color impression of an original image. Resultant images obtained by the color transfer methods are appropriately mixed into an output image. In experiments, the proposed method is applied to a variety of images and the effectiveness of the proposed method is confirmed by subjective and objective evaluations.

Recently, various low-light image enhancement methods are proposed. LIME is one of the low-light image enhancement methods based on Retinex theory. Retinex theory assumes that the image is decomposed into the illumination component and the reflectance component. In LIME, the illumination map is estimated by solving optimization problem for edge-aware smoothing and the output image is obtained by adjusting the illumination map. Though LIME significantly enhances the contrast in dark regions, it tends to cause clipping in the bright regions. In this paper, we propose a new method of low-light image enhancement. In the proposed method, firstly the illumination map is estimated by the soft-closing operation using smoothed local histogram. Then, the output image is obtained by adjusting the illumination in the same way as LIME. The proposed method can improve the contrast of the image overall while suppressing clipping in bright regions. The effectiveness of the proposed method is verified by experiments using various images.

By examining the state of the retinal blood vessels, medical doctors can diagnose the degree of arteriosclerosis of retinal blood vessels. However, its diagnosis is highly dependent on the experience of medical doctors. We had previously proposed a method that could evaluate quantitatively the degree of arteriosclerosis of the retinal blood vessels. In that method, the contour lines of vein were firstly extracted using energy function, and then the caliber of vein was calculated. However, this conventional method sometimes extracted the unexpected contour lines of blood vessels, and the points to calculate the caliber of vein for evaluation of the degree of arteriosclerosis had to be given manually. In this study, we propose an automatic method for evaluation of arteriosclerosis without any manual operations. In the experiments, the proposed method has been successfully applied to 400 actual eye fundus images, and the experimental results have shown that the proposed method is promising.

Coronary atherosclerotic plaque has been extensively studied in pathological research. Improving the evaluation of vulnerable rupture is important to prevent acute heart failure. Intravascular ultrasound (IVUS) method is one of techniques to acquire information about atherosclerotic plaque, which is backscattered ultrasound signal sensed by IVUS transducer. The vessel structure and tissue components are then characterized in relation to the acquired signals. In this study, eight human coronary vessel sections are involved, and we use discriminative restricted Boltzmann machine (RBM) to classify coronary tissues as a target classifier. The quantization domain of IVUS signals are used to extract binary features for adapting Gaussian model of RBM. In addition, we propose a misclassification sensitive training of disRBM to deal with the class imbalances. The results are compared to the conventional integrated backscattered IVUS method (IB-IVUS) and the cost sensitive neural network for the same tasks.

High performance of color quantization processing is very important for obtaining limited-color images with good quality. The median cut algorithm (MCA) is a typical color quantization method. Its computational cost is low owing to its simple algorithm, but the quality of output images is mediocre at best. In this paper, we describe a modification of MCA. In our method, we use a combination of principal component analysis (PCA) and linear discriminant analysis (LDA) to accomplish effective partitioning of color space. Concretely, PCA and LDA are used to calculate partitioning planes and their positions, respectively. We verify the effectiveness of our method through experiments using 24-bit full-color natural images.

In the present paper, we propose an image contrast enhancement method that can enhance the contrast of a color image naturally by taking account of a color space shape. The proposed method realizes the natural enhancement based on two kinds of intensity histograms: a gradient-norm-based histogram and an ideal histogram derived from the shape of a color space. The former histogram is used to suppress over-enhancement in the flat regions of an image and the latter histogram is used to prevent the whole image from being darken. Concretely, the aforementioned intensity histograms are appropriately mixed into a histogram with a weight based on the average intensity of the input image. The contrast enhancement of the input image is realized using the cumulative histogram of the mixed histogram as an intensity transform function. To verify the validity of the proposed method, in experiments, the proposed method is applied to a variety of images and experimental results are evaluated qualitatively and quantitatively.

A novel color removal method for digital color images is proposed. In this method, differences of colors in an input color image are reflected effectively, and the perceived lightness coming from the input color image is preserved as much as possible in order to realize the natural color-to-gray conversion. The effectiveness and the validity of the proposed method are verified by some experiments.

The color appearance of a food image is very important for making it look delicious. Many food images acquired in low illumination environments do not give delicious impressions because of the deterioration of their color appearance. We propose an image enhancement method to improve the color appearance of a food image acquired in a low illumination environment. In the proposed method, the lightness of each pixel is first enhanced within a color gamut on an equal hue plane in the CIELAB color space. The boundary of the color gamut is given by a lookup table calculated in advance. Then, when the hue of the pixel is attributed to foods, the chroma is enhanced within the color gamut. In this case, we consider Helmholtz–Kohlrausch (H-K) effect, which indicates the relation of the hue, chroma, and lightness in human visual perception. According to H-K effect, the perceived lightness increases as the chroma is enhanced. Therefore, the perceived lightness is preserved after the chroma enhancement, so that the food image is not too enhanced. To verify the effectiveness of the proposed method, experiments are conducted using several food images.

This paper describes a novel image filter with superior performance on detail-preserving removal of random-valued impulse noise superimposed on natural gray-scale images. The non-local means filter is in the limelight as a way of Gaussian noise removal with superior performance on detail preservation. By referring the fundamental concept of the non-local means, we had proposed a non-local median filter as a specialized way for random-valued impulse noise removal so far. In the non-local processing, the output of a filter is calculated from pixels in blocks which are similar to the block centered at a pixel of interest. As a result, aggressive noise removal is conducted without destroying the detailed structures in an original image. However, the performance of non-local processing decreases enormously in the case of high noise occurrence probability. A cause of this problem is that the superimposed noise disturbs accurate calculation of the similarity between the blocks. To cope with this problem, we propose an improved non-local median filter which is robust to the high level of corruption by introducing a new similarity measure considering possibility of being the original signal. The effectiveness and validity of the proposed method are verified in a series of experiments using natural gray-scale images.

The performance of human visual functions degrades with age. For example, the visual field becomes dark owing to a decrease of the amount of light received. Moreover, objects appear to be hazy by a decrease in the contrast sensitivity. In particular, the contrast sensitivity decreases remarkably as the spatial frequency increases. In this paper, we propose an image enhancement method which consists of a lightness conversion and an amplification of high spatial frequency components to take account of visual characteristics of elderly persons. Concretely, the lightness conversion is conducted within the color gamut by using a tone mapping. In addition, a sharpening based on the inverse characteristic function of the sensitivity decay curve is applied. The effectiveness of the proposed method is verified by experiments using several digital images.

The dynamic range of human light perception for natural scenes is wider than the dynamic ranges of cameras. When an image of a natural scene is taken by a camera, the image may contain under-exposed or over-exposed regions. A remedy for this problem is to fuse multi-exposure images which are taken by the camera at different exposure levels. Since most fusion methods use a weighted average of multi-exposure images, the appropriate calculation of its weights is important. In this paper, we propose a method for fusing multi-exposure images which are newly generated by a light conversion with preserving hue and saturation. In the proposed method, the lightness conversion is based on local color correction (LCC) and a fused image is obtained by a weighted average whose weights are calculated based on saturation. The proposed method is applied to several images and the effectiveness of the proposed method is investigated.

The color of an object appears to be different for elderly and young persons. This is because human visual features vary with on age. In general, the spectral transmittance of the crystalline lens of an elderly person is lower than that of a young person. The low spectral transmittance decreases the brightness of the visual field. This phenomenon affects the color appearance of an image in human visual perception. In this paper, a lightness transform method is proposed to improve the color appearance of images for elderly persons. In the proposed method, colors that are difficult for an elderly person to see are transformed into colors that are easy to see by adding lightness contrast. The performance of the proposed method is confirmed in experiments using digital images.

Graphical abstractDisplay Omitted HighlightsThis work proposes new IVUS-based tissue characterization method that uses a mnSOM.The proposed method can create a map of various dynamical features from the radio-frequency signal.The proposed method has a high generalization performance for IVUS -based tissue characterization as compared with conventional method. The intravascular ultrasound-based tissue characterization of coronary plaque is important for the early diagnosis of acute coronary syndromes. The conventional tissue characterization techniques however cannot obtain sufficient identification accuracy for various tissue properties, because the feature employed for characterization are static features, which lack dynamical information about backscattered radio-frequency (RF) signals.In this work, we propose a new intravascular ultrasound-based tissue characterization method that uses a modular network self-organizing map (mnSOM) in which each module is composed of an autoregressive model for representing the dynamics of the RF signals.The proposed method can create a map of various dynamical features from the RF signal. This map enables generalized tissue characterizations. The proposed method is verified by comparing its tissue characterization performance with that of the conventional method using real intravascular ultrasound signals.

This paper describes a novel image filtering method that removes random-valued impulse noise superimposed on a natural color image. In impulse noise removal, it is essential to employ a switching-type filtering method, as used in the well-known switching median filter, to preserve the detail of an original image with good quality. In color image filtering, it is generally preferable to deal with the red (R), green (G), and blue (B) components of each pixel of a color image as elements of a vectorized signal, as in the well-known vector median filter, rather than as component-wise signals to prevent a color shift after filtering. By taking these fundamentals into consideration, we propose a switching-type vector median filter with non-local processing that mainly consists of a noise detector and a noise removal filter. Concretely, we propose a noise detector that proactively detects noise-corrupted pixels by focusing attention on the isolation tendencies of pixels of interest not in an input image but in difference images between RGB components. Furthermore, as the noise removal filter, we propose an extended version of the non-local median filter, we proposed previously for grayscale image processing, named the non-local vector median filter, which is designed for color image processing. The proposed method realizes a superior balance between the preservation of detail and impulse noise removal by proactive noise detection and non-local switching vector median filtering, respectively. The effectiveness and validity of the proposed method are verified in a series of experiments using natural color images.

Generally, in color image contrast enhancement, the intensity or saturation component of an image is processed without the change of hue component of the image. In some color space used for contrast enhancement, processed pixel values are located outside of the gamut of the RGB color space. In that case, a clipping process is required to make the pixel values located inside of the gamut. However, the clipping process changes the combination of R, G, and B components and causes unexpected hue changes. In the present paper, we propose a hue-preserving contrast enhancement method for color images without gamut problem. In the proposed method, the contrast enhancement of a color image is realized by spreading out the distribution of pixel values on a constant hue plane in the RGB color space. The constant hue plane is represented as the triangle whose vertices correspond to white, black, and the maximally saturated colors. By spreading out the distribution of the pixel values on the constant hue plane, a remedy for the gamut problem and hue-preserving contrast enhancement are realized. Through experiments, the effectiveness of the proposed method is verified.

White balance (WB) is the correction process to remove unrealistic color casts caused by an illuminant color. For high-quality image correction, it is important to accurately estimate the illuminant color. In this study, we propose an accurate illuminant color estimation method for WB based on pseudo-detection of specular reflection. In the proposed method, pixels which are brighter than their neighborhood and their neighboring darker pixels are found at first. Then, the averages of their pixel values are calculated respectively, and then the specular reflection is detected approximately as the difference of the average values. Finally, the illuminant color is estimated as the specular reflection itself. The accurate estimation capability of the proposed method is verified through a series of comparative experiments with typical WB methods.

This paper describes a novel image filtering method for removal of random-valued impulse noise superimposed on grayscale images. Generally, it is well known that switching-type median filters are effective for impulse noise removal. In this paper, we propose a more sophisticated switching-type impulse noise removal method in terms of detail-preserving performance. Specifically, the noise detector of the proposed method finds out noise-corrupted pixels by focusing attention on the difference between the value of a pixel of interest (POI) and the median of its neighboring pixel values, and on the POI’s isolation tendency from the surrounding pixels. Furthermore, the removal of the detected noise is performed by the newly proposed median filter based on non-local processing, which has superior detail-preservation capability compared to the conventional median filter. The effectiveness and the validity of the proposed method are verified by some experiments using natural grayscale images.

This paper proposes a hybrid boundary detection method for image based on a new modified level set method and a fuzzy model. It is applied to a boundary detection problem of coronary plaque. Level set method has been applied widely in image processing. It however does not work well for an intravascular ultrasound (IVUS) image because an image gradient, commonly used for calculating a speed function in the level set method, cannot detect an image boundary well. The level set method and the weighted image separability proposed by the authors in the past were applied for a coronary plaque boundary detection problem. The level set method could not however detect the plaque boundary in several regions. One problem was that the candidates of the plaque boundary detected by the weighted separability were unclear in several regions. The other problem was that the IVUS image often becomes shadowed and it contains no texture information there, which is caused by the presence of the guide wire. To overcome this problem, we propose a new modified level set, and we further propose a hybrid boundary detection method based on the new modified level set and the Takagi Sugeno (T-S) fuzzy model for detecting a coronary plaque boundary. The boundary detection accuracy of the proposed method was significantly better than those of the previous methods we proposed in the past.

Conventionally, in order to sharpen and enhance blurred digital images, unsharp masking (UM) has been used widely. However, over-enhancement is generated frequently in edge parts of the image and the artifacts such as jaggy and halo are sometimes seen in the resultant image. To cope with the problem, the selective UM methods were proposed so far. However, they were not always effective for the realization of the artifact suppression and image sharpening at the same time. In this paper, we propose new selective UM methods which employ the bilateral filter and Laplacian of intensity and/or saturation components of image. In the methods, the occurrence of jaggy and halo is suppressed by weakening the sharpening effect especially at the edge parts of objects in the image, and the image is well sharpened totally. Through the experiments using some images, the effectiveness of the proposed methods is illustrated. 

In comparison with the people having the normal color vision, the people with color vision impairment such as dichromats have the difficulty to distinguish the colors in certain combinations. This may pose an impediment of the visual communication in the recent wide spreading of the use of colors. In order to address this problem, in this paper, we propose a novel lightness modification method to improve the visibility of the color image for the color vision impaired. In the proposed method, the lightness modification is only carried out around contours of objects to realize the Craik–O'Brien effect, which is an perceptual illusion of the contrast, for dichromats. Furthermore, in the proposed method, it is tried to preserve the impression given by original color image by lessening the change of the color as possible while providing the perceptible and comprehensible color image for dichromats. Through some examples, the effectiveness and the validity of the proposed method are verified.

It is known that the lenticular transmittance of elderly person is lower than that of young person. This low lenticular transmittance property causes the brightness decrease of the visual field of the elderly person. In this report, the lightness transform method of the image for the elderly person is proposed. In the proposed method, the colors, which are hard to see for the elderly person, are transformed into the colors, which are easy to see, by adding them the lightness contrast. The effectiveness of the proposed method is verified by some experiments.

This paper describes a novel image filtering method for removal of random-valued impulse noise superimposed on color images. In color image filtering, generally, it is preferable to deal with the R, G, and B components of each pixel of a color image as elements of a vectorized signal as in the vector median filter rather than component-wise signals in order to prevent color shift after filtering. Furthermore, in impulse noise removal, it is essential to use switching type filtering manner as in the switching median filter to preserve the detailed part of an original image with good quality. By taking those fundamentals in consideration, in this study, we propose a switching-type vector median filter with non-local processing mainly consists of a noise detection part and a noise removal one. Concretely, we propose a noise detector which proactively finds out noise-corrupted pixels by focusing attention on isolation tendencies of pixels of interest in difference images between RGB components. Furthermore, as a noise removal filter, we propose an extended version of Non-local median filter (NL-MF), which we proposed previously for grayscale images, named Non-local vector median filter (NL-VMF) which deals with color images. The proposed method realizes superior balance between the detail-preservation based on the proactive noise detection and the noise removal by the non-local switching vector median filtering. The effectiveness and the validity of the proposed method are verified by some experiments using natural color images.

In this paper, a new image enlargement method applying the backprojection for lost pixel (BPLP) to the predefined codebook-based method is proposed. BPLP is a method for image restoration. In BPLP, the eigenspace reflecting the characteristics of an input image is generated from the remained pixels and is used to restore the missing pixels. In the proposed method, the eigenspace is replaced by one generated from the predefined codebook (PDC). PDC represents edge-blurring properties in a small image patch and consists of pairs of low- and high-frequency image patches on various edge patterns. By replacing the PDC-based estimation of lost high-frequency components with BPLP, a fast image enlargement method retaining its performance can be developed. Through some experiments, the effectiveness of the proposed method was demonstrated. Especially, it was confirmed that the processing time of the proposed method was shortened to about 1/50 that of the PDC-based method.

A new recoloring method to improve visibility of indiscriminable colors for protanopes or deuteranopes is proposed. In the proposed method, yellow-blue components of a color image perceived by protanopes/deuteranopes are adequately modified. Moreover, the gamut mapping is considered to obtain proper output color values in this method.

Color transfer involves recoloring reflecting a reference image color in an input image, for example, by a color distribution or chroma distribution. The color transfer algorithm we propose transferring hue information by modifying the input image hue histogram so that its shape approaches that of the reference image as closely as possible. Color transfer effectively reflecting images considers image chroma information and relative hue histogram occurrence frequency as demonstrated in results of experiments.

The novel image enhancement approach we propose is based on blending the ε-filtering and a conventional unsharp masking. The proposed selective unsharp masking is especially effective in enhancing noisy images corrupted by Gaussian noise, as confirmed in image enhancement experiments.

In order to convert a color image into a monochrome one, the lightness components of pixels have to date been used as gray-levels for the representation of the monochrome image. However, saliencies of an image embedded only in the chrominance components are eliminated in such conversion. To cope with this problem, “Color2Gray” algorithm, which excels in the color removal of digital images, has been proposed by Gooch et al. [ACM Trans. Graphics 24 (2005) 634]. In this paper, the algorithm is first analyzed and its mathematical property is revealed. Then a fast Color2Gray algorithm is proposed by using the mathematical property. Finally, the validity and the effectiveness of the proposed algorithm are proven by some experiments.

A new image super-resolution method based on a codebook mapping is proposed. The codebook mapping represents the internal relationship between low-and high-frequency image components, and is used for the estimation of highfrequency image components lost in the sampling process. In the proposed method, codebooks are first generated by the low-and high-frequency image components of the original image itself that is to be processed. Then, a resultant superresolution image, that is, an enlarged image, is obtained by combining the estimated high-frequency image components with the low-frequency ones obtained by a typical interpolation-based method. The effectiveness of the proposed method is verified by comparing it against some conventional methods.

A novel hierarchical k-nearest neighbor classification method using the feature and observation space information is proposed. The present method performs a fine classification when a pair of the spatial coordinate of the observation data in the observation space and its corresponding feature vector in the feature space is provided.

This paper describes a new type of self-organizing map (SOM) with twin units as opposed to the single unit type conventional SOM proposed by Kohonen. The present self-organizing map with twin units (TW-SOM) can describe a nonlinear input-output relation with high accuracy. It is applied to voice conversion problem from bone conduction voice to air conduction voice (nonlinear code vector mapping), and its superiority over the conventional method using Linde-Buzo-Gray (LBG) algorithm is discussed. The tone quality of the converted voice is examined not only from the quantization distortion viewpoint, but also from the auditory sensation viewpoint through actual listening tests. The enhancement of the tone quality was experimentally confirmed.

The estimation of the point spread function (PSF) is a very important and indispensable task for practical image restoration. Various PSF estimation algorithms have been developed, especially for the out-of-focus blur. However, a majority of them are useless in an extremely noisy environment. This paper describes a new robust PSF estimation algorithm based on a distribution of gradient vectors on the logarithmic amplitude spectrum mapped to the polar plane. The proposed algorithm can estimate the out-of-focus PSF accurately and robustly, even for an image highly corrupted by noise. The effectiveness of the proposed algorithm is verified by applying it to the PSF estimation for out-of-focus blurred and noisy images.

Intelligent scissors is an interactive image segmentation algorithm which allows a user to select piece-wise globally optimal contour segment corresponding to a desired object boundary. However, the intelligent scissors is too sensitive to a noise and texture patterns in an image since it utilizes the gradient information concerning the pixel intensities. This paper describes a new intelligent scissors based on the concept of the separability in order to improve the object boundary extraction performance. The effectiveness of the proposed method has been confirmed by some experiments for actual images acquired by an ordinary digital camera.

The estimation of the point-spread function (PSF) is one of very important and indispensable tasks for the practical image restoration. Especially, for the motion blur, various PSF estimation algorithms have been developed so far. However, a majority of them becomes useless in the low blurred signal-to-noise ratio (BSNR) environment. This paper describes a new robust PSF estimation algorithm based on Hough transform concerning gradient vectors, which can accurately and robustly estimate the motion blur PSF even in low BSNR case. The effectiveness and validity of the proposed algorithm are verified by applying it to the PSF estimation and the image restoration for noisy and motion blurred images.

It is known that the lenticular transmittance of elderly people is lower than that of young people. Especially, it is remarkable for short wavelength lights. Therefore blue-tinged colors become darker comparing with the other colors in the field of vision of elderly people. In this paper, in order to improve the visibilities of blue-tinged colors in the elderly vision, the data-dependent gamma correction is proposed. Concretely, RGB components are converted by the gamma correction with a gamma which is set according to the hue, saturation and lightness contrast in each pixel. Through the experiments using some digital color images, the effectiveness of the proposed method is verified.

This paper proposes a new blind deconvolution method with additional phase spectral constraints for a blurred image. A degradation of an original image is mathematically modeled by a convolution of an original image and a point-spread function (PSF). The proposed method consists of the following three steps: (i) projection onto a complex set satisfying the phase spectral constraint in a frequency space; (ii) minimization of a cost function preserving the constrained phase spectra; and (iii) projection onto an image space satisfying nonnegative and support constraints. This method restores both the original image and the PSF with high accuracy. The effectiveness of the proposed method is verified by applying it to some blind deconvolution problems for digital images, and the experimental results show that the performance is superior to the conventional blind deconvolution methods.