This paper introduces an automatic and multi-instrument snowfall observation system and proposes techniques that could be used in the estimation of snowfall characteristics. The instruments used in this study include two microwave radars, an optical lidar, a CCD camera based imaging system and high-accuracy electrical balances for reference data. The emphasis has been on obtaining good temporal resolution and synchronization accuracy of separate datasets. In most research done so far, this has not been a principal point, either because only very long snowfall events have been measured, or wide area estimates were desired, or due to limitations in manual sampling methods and other technical issues. The measurements were also contained in a small area to make sure that all instruments record data from the same target. One radar and the optical lidar recorded an atmospheric profile up to 6000 m, while the other radar, the imaging system and the two balances recorded snowfall on the ground level. The combination of optical, microwave and direct visual observations of snowfall show that a change in cloud conditions can result in snowfall having different characteristics. The lidar backscatter was used as main indicator of transitions in cloud conditions. A direct visual evaluation of snowflake size distribution using a CCD camera shows that it is extremely helpful in order to interpret radar data. The camera observed velocity distribution showed no large variations between snowfall events, however, it could be useful in detecting graupel and hail precipitations which have much faster terminal velocities. This paper will conclude with a discussion on further elaborating the use of lidar and visual data to complement radar observations of snowfall.

This paper presents a method for detection of calcification, which is an important early sign of breast cancer in mammograms. Since information of calcifications is located in inhomogeneous background and noises, it is hard to be detected. This method uses wavelet packet transform (WPT) for elimination of the background image related to low frequency components. However, very high frequency signals of noises exist with the calcifications and make it hard to suppress them. Since calcification location can be represented as vertical, horizontal, and diagonal edges in time-frequency domain, the edges in spatial domain can be utilized as a filter for noise suppression. Then the image from inverse transform will contain only required information. A free-response operating characteristic (FROC) curve is used to evaluate a performance of proposed method by applying it to thirty images of calcifications. The results show 82.19 percent true positive detection rate at the cost of 6.73 false positive per image.

Recently, it becomes popular to synthesize new viewpoint images based on some sampled viewpoint images of real scene using technique of computer vision. 3D shape reconstruction in Euclidean space is not necessarily required, but information of dense matching points is basically enough to synthesize new viewpoint images. In this paper, we propose a new method for 3D reconstruction from three cameras based on projective geometry. In the proposed method, three input camera images are rectified based on projective geometry, so that the vertical and horizontal directions can be completely aligned with the epipolar planes between the cameras. This rectification provides Projective Voxel Space (PVS), in which the three axes are aligned with the directions of camera projection. Such alignment simplifies the procedure for projection between the 3D space and the image planes in PVS. Taking this advantage of PVS, silhouettes of the objects are projected into PVS, so that the searching area of matching points can be reduced. The consistency of color value between the images is also evaluated for final determination of the matching point. The finally acquired matching points in the proposed method are described as the surface of the objects in PVS. The acquired surface of the objects in PVS also includes knowledge about occlusion. Finally, images from new viewpoints can be synthesized from the matching points and occlusions. Although the proposed method requires only weak calibration, plausible occlusions are also synthesized in the images. In the experiments, images of virtual viewpoints, which were set among three cameras, are synthesized from three real images.

This paper addresses the problem to extract moving object from the moving background in the mixed domain (MixeD), which makes it possible to carry the filtering in one dimension. Since the velocities of moving object and background are necessary for moving object extraction, we first estimate the velocities based on the appropriate spatial frequency point selection method in the MixeD. Then an optimal filter used for 1-D signal filtering is designed. By filtering 1-D signals over all spatial frequencies, signals with certain velocity vector are extracted, while the extracted image is obtained by applying the 2-D IDFT to the filtered signals. The simulation results show that the method can extract moving object based both on the correctly estimated velocity and the proposed optimal 1-D filter.

This paper describes the implementation of a proposed image filter into a Discrete-Time Cellular Neural Network (DT-CNN). The three stages that compose the filter are described, showing that the resultant filter is capable of (1) erasing or detecting several concentric shapes simultaneously, (2) thresholding and (3) thinning of gray-scale images. Because the DT-CNN has to fill certain conditions for this filter to be implemented, it becomes a modified version of a DT-CNN. Those conditions are described and also experimental results are clearly shown.

Processes in image recognition include target detection and shape extraction. Active Net has been proposed as one of the methods for such processing. It treats the target detection in an image as an energy optimization problem. In this paper, a problem of the conventional Active Net is presented and the new Active Net is proposed. The new net is improved the ability for detecting a target. Finally, the validity of the proposed net is confirmed by experimental results.

A new adaptive multichannel filtering approach is introduced and analyzed in this paper. The technique is simpler and more appropriate than traditional approaches that have been addressed by means of groupwise vector ordering information. These filters are a two-stage filters based on rational functions (RF) using fuzzy transformations of the Euclidean and angular distances among the different vectors to adapt to local data in the color image. The output is the result of vector rational operation taking into account three fuzzy sub-function outputs. Simulation studies indicate that the filters are computationally attractive and have excellent performance such as edge and details preservation and accurate chromaticity estimation.

The classified side-match vector quantizer, CSMVQ, has already been presented for low-bit-rate image encoding. It exploits a block classifier to decide which class the input vector belongs to using the variances of the upper and left codewords. However, this block classifier doesn't take the variance of the current input vector itself into account. This letter presents a new CSMVQ in which a two-level block classifier is used to classify input vectors and two different master codebooks are used for generating the state codebook according to the variance of the input vector. Experimental results prove the effectiveness of the proposed CSMVQ.

An averaged intensity peak profile of light scattered from a random medium depends on a thickness of a sample as well as parameters such as a volume fraction and a size of particles composing the medium. We used this dependence to measure a depth profile varied in the random medium. We demonstrated the possible simultaneous measurement of a transport mean free path and a depth of an aqueous suspension of titanium particles.

To estimate moving object velocity in an image sequence is useful for a variety of applications, such as velocity measurement, computer vision and monitoring systems. An effective way is to approach it in the transform/spatiotemporal mixed domain (MixeD), which transforms the 3-D signal processing problem into 1-D complex signal processing. But it remains a problem how to select several spatial frequency points in the MixeD which may influence the accuracy of velocity estimation and object detection. In this paper, a spatial frequency selection method has been proposed, which can choose the appropriate spatial frequency points out of a number of points in MixeD automatically. So the velocity estimation problem can be addressed by solving the coupled equations established over two selected appropriate points in 2-D spatial frequency domain other than searching for the spectral energy plane over a number of points selected by experience. In this method, evaluation functions corresponding to image sequence with one moving object and two moving objects are established firstly, and the selection is then achieved by using the established evaluation functions together with a threshold. The simulation results show that the proposed method is effective on the appropriate spatial frequency selection.

This paper presents a novel competitive learning algorithm for the design of variable-rate vector quantizers (VQs). The algorithm, termed variable-rate competitive learning (VRCL) algorithm, designs a VQ having minimum average distortion subject to a rate constraint. The VRCL performs the weight vector training in the wavelet domain so that required training time is short. In addition, the algorithm enjoys a better rate-distortion performance than that of other existing VQ design algorithms and competitive learning algorithms. The learning algorithm is also more insensitive to the selection of initial codewords as compared with existing design algorithms. Therefore, the VRCL algorithm can be an effective alternative to the existing variable-rate VQ design algorithms for the applications of signal compression.

In this paper, we describe a method of determining meat quality using the concepts of "marbling score" and texture analysis. The marbling score is a measure of the density distribution of fat in the rib-eye region. Based on the results of an investigation carried out by handing out questionnaires to graders, we consider the marbling of meat to be a texture pattern and propose a method for the implementation of a grading system using a texture feature. In this system, we use a gray level co-occurrence matrix as the texture feature, which is a typical second-order statistic of gray levels of a texture image, and determine standard texture-feature vectors for each grade based on it. The grade of an unevaluated image is determined by comparing the texture-feature vector of this unevaluated image with the standard texture-feature vectors. Experimental results show the proposed method to be effective.

A repeating watermarking technique based on visual secret sharing (VSS) scheme provides the watermark repeated throughout the image for avoiding the image cropping. In this paper, the watermark is divided into public watermark and secret watermark by using the VSS scheme to improve the security of the proposed watermarking technique. Unlike the traditional methods, the original watermark does not have to be embedded into the host image directly and, thus, it is hard to be detected or removed by the pirates or hackers. The retrieved watermark extracted from the watermarked image does not require the complete original image, but requires a secret watermark. Furthermore, the watermarking technique suits the watermark with an adaptive size of binary image for designing the watermarking system. The experimental results show that the proposed method can withstand the common image processing operations, such as filtering, lossy compression and the cropping attacking etc. The embedded watermark is imperceptible, and that the extracted watermark identifies clearly the owner's copyright.

A high speed 3D shape reconstruction method with multiple video cameras and multiple computers on LAN is presented. The video cameras are set to surround the real 3D space where people exist. Reconstructed 3D space is displayed in voxel format and users can see the space from any viewpoint with a VR viewer. We implemented a prototype system that can work out the 3D reconstruction with the speed of 10.55 fps in 313 ms delay.

Identification of motion parameters is an important issue in image restoration of a linear motion blur. Based on the human visual-motion sensing properties, an integrated approach with some known image processing techniques is proposed to the estimation of the direction and extent of motion on a linear motion blurred image. Experimental results confirm the feasibility of our approach.

The visual inspection of printed circuit boards (PCBs) at the final production stage is necessary for quality assurance and the requirements for an automated inspection system are very high. However, consistent inspection of patterns on these PCBs is very difficult due to pattern complexity. Most of the previously developed techniques are not sensitive enough to detect defects in complex patterns. To solve this problem, we propose a new optical system that discriminates pattern types existing on a PCB, such as copper, solder resist and silk-screen printing. We have also developed a hybrid defect detection technique to inspect discriminated patterns. This technique is based on shape measurement and features extraction methods. We used the proposed techniques in an actual automated inspection system, realizing real time transactions with a combination of hardware equipped with image processing LSIs and PC software. Evaluation with this inspection system ensures a 100% defect detection rate and a fairly low false alarm rate (0.06%). The present paper describes the inspection algorithm and briefly explains the automated inspection system.

Fuzzy rule-based edge detection using multiscale edge images is proposed. In this method, the edge image is obtained by fuzzy approximate reasoning from multiscale edge images which are obtained by derivative operators with various window sizes. The effect of utilizing multiscale edge images for edge detection is already known, but how to design the rules for deciding edges from multiscale edge images is not clarified yet. In this paper, the rules are represented in a fuzzy style, since edges are usually defined ambiguously, and the fuzzy rules are designed optimally by a training method. Here, the fuzzy approximate reasoning is expressed as a nonlinear function of the multiscale edge image data, and the nonlinear function is optimized so that the mean square error of the edge detection be the minimum. Computer simulations verify its high performance for actual images.

Watermarking methods that employ orthogonal transformations are very robust against non-geometrical modifications such as lossy compression, but attaining robustness against image translation or cropping is difficult. This report describes a watermarking method that increases robustness against geometrical modifications such as image translation and cropping by embedding watermark data in the frequency component of an image and detecting that data by considering the phase difference of the coefficients that results from translation of the image. Experimental results demonstrate the robustness of this method against both non-geometrical image changes and image translation and cropping.

Image coding with vector quantization (VQ) reveals several defects which include edge degradation and high encoding complexity. This paper presents an edge-preserving coding system based on VQ to overcome these defects. A signal processing unit first classifies image blocks into low-activity or high-activity class. A high-activity block is then decomposed into a smoothing factor, a bit-plane and a smoother (lower variance) block. These outputs can be more efficiently encoded by VQ with lower distortion. A set of visual patterns is used to encode the bit-planes by binary vector quantization. We also develop a modified search-order coding to further reduce the redundancy of quantization indexes. Simulation results show that the proposed algorithm achieves much better perceptual quality with higher compression ratio and significant lower computational complexity, as compared to the direct VQ.

This paper proposes applications of a code-division multiplexing technique to VLSI systems free from interconnection problems. We employ a pseudo-random orthogonal m-sequence carrier as a multiplexable information carrier to achieve efficient data transmission. Using orthogonal property of m-sequences, we can multiplex several computational activities into a single circuit, and execute in parallel using multiplexed data transmission with reduced interconnection. Also, randomness of m-sequences offers the high tolerance to interference (jamming), and suppression of dynamic range of signals while maintaining a sufficient signal-to-noise ratio (SNR). We demonstrate application examples of multiplex computing circuits, neural networks, and spread-spectrum image processing to show the advantages.