This PDF file contains the front matter associated with SPIE Proceedings Volume 9814 including the Title Page, Copyright information, Table of Contents, Introduction, Authors, and Conference Committee listing.

Ant colony algorithm is heuristic, bionic and parallel. Because of it is property of positive feedback, parallelism and simplicity to cooperate with other method, it is widely adopted in planning on discrete space. But it is still not good at planning on continuous space. After a basic introduction to the basic ant colony algorithm, we will propose an ant colony algorithm on continuous space. Our method makes use of the following three tricks. We search for the next nodes of the route according to fixed-step to guarantee the continuity of solution. When storing pheromone, it discretizes field of pheromone, clusters states and sums up the values of pheromone of these states. When updating pheromone, it makes good resolutions measured in relative score functions leave more pheromone, so that ant colony algorithm can find a sub-optimal solution in shorter time. The simulated experiment shows that our ant colony algorithm can find sub-optimal solution in relatively shorter time.

The serious information redundancy in hyperspectral images (HIs) cannot contribute to the data analysis accuracy, instead it require expensive computational resources. Consequently, to identify the most useful and valuable information from the HIs, thereby improve the accuracy of data analysis, this paper proposed a novel hyperspectral band selection method using the hybrid genetic algorithm and gravitational search algorithm (GA-GSA). In the proposed method, the GA-GSA is mapped to the binary space at first. Then, the accuracy of the support vector machine (SVM) classifier and the number of selected spectral bands are utilized to measure the discriminative capability of the band subset. Finally, the band subset with the smallest number of spectral bands as well as covers the most useful and valuable information is obtained. To verify the effectiveness of the proposed method, studies conducted on an AVIRIS image against two recently proposed state-of-the-art GSA variants are presented. The experimental results revealed the superiority of the proposed method and indicated that the method can indeed considerably reduce data storage costs and efficiently identify the band subset with stable and high classification precision.

For the real-time motion deblurring, it is of utmost importance to get a higher processing speed with about the same image quality. This paper presents a fast Richardson-Lucy motion deblurring approach to remove motion blur which rotates blurred image under blurring paths. Hence, the computational time is reduced sharply by using one-dimensional Fast Fourier Transform in one-dimensional Richardson-Lucy method. In order to obtain accurate transformational results, interpolation method is incorporated to fetch the gray values. Experiment results demonstrate that the proposed approach is efficient and effective to reduce motion blur under the blur paths.

Bidimensional convolution is a low-level processing algorithm of interest in many areas, but its high computational cost constrains the size of the kernels, especially in real-time embedded systems. This paper presents a hardware architecture for the ASIC-based implementation of 2-D convolution with medium–large kernels. Aiming to improve the efficiency of storage resources on-chip, reducing off-chip bandwidth of these two issues, proposed construction of a data cache reuse. Multi-block SPRAM to cross cached images and the on-chip ping-pong operation takes full advantage of the data convolution calculation reuse, design a new ASIC data scheduling scheme and overall architecture. Experimental results show that the structure can achieve 40× 32 size of template real-time convolution operations, and improve the utilization of on-chip memory bandwidth and on-chip memory resources, the experimental results show that the structure satisfies the conditions to maximize data throughput output , reducing the need for off-chip memory bandwidth.

Discrete Fourier transform (DFT) is one of the most wildly used tools for signal processing. In this paper, a novel sliding window algorithm is presented for fast computing 2D DFT when sliding window shifts more than one-point. The propose algorithm computing the DFT of the current window using that of the previous window. For fast computation, we take advantage of the recursive process of 2D SDFT and butterfly-based algorithm. So it can be directly applied to 2D signal processing. The theoretical analysis shows that the computational complexity is equal to 2D SDFT when one sample comes into current window. As well, the number of additions and multiplications of our proposed algorithm are less than those of 2D vector radix FFT when sliding window shifts mutiple-point.

This paper presents hardware efficient designs for implementing the one-dimensional (1D) discrete Fourier transform (DFT). Once DFT is formulated as the cyclic convolution form, the improved first-order moments-based cyclic convolution structure can be used as the basic computing unit for the DFT computation, which only contains a control module, a barrel shifter and (N-1)/2 accumulation units. After decomposing and reordering the twiddle factors, all that remains to do is shifting the input data sequence and accumulating them under the control of the statistical results on the twiddle factors. The whole calculation process only contains shift operations and additions with no need for multipliers and large memory. Compared with the previous first-order moments-based structure for DFT, the proposed designs have the advantages of less hardware consumption, lower power consumption and the flexibility to achieve better performance in certain cases. A series of experiments have proven the high performance of the proposed designs in terms of the area time product and power consumption. Similar efficient designs can be obtained for other computations, such as DCT/IDCT, DST/IDST, digital filter and correlation by transforming them into the forms of the first-order moments based cyclic convolution.

A method of implement FFT based on FPGA IP Core is introduced in this paper. In addition, for the spectrum leakage caused by the truncation of the non-integer-period sampling, an improved method of adding window to the input signal to restrain the spectrum leakage is proposed. The design was simulated in the Matlab environment. The results show that the proposed method has good performance with some improvement.

This paper realizes the automatic segmentation and classification of Mycobacterium tuberculosis with conventional light microscopy. First, the candidate bacillus objects are segmented by the marker-based watershed transform. The markers are obtained by an adaptive threshold segmentation based on the adaptive scale Gaussian filter. The scale of the Gaussian filter is determined according to the color model of the bacillus objects. Then the candidate objects are extracted integrally after region merging and contaminations elimination. Second, the shape features of the bacillus objects are characterized by the Hu moments, compactness, eccentricity, and roughness, which are used to classify the single, touching and non-bacillus objects. We evaluated the logistic regression, random forest, and intersection kernel support vector machines classifiers in classifying the bacillus objects respectively. Experimental results demonstrate that the proposed method yields to high robustness and accuracy. The logistic regression classifier performs best with an accuracy of 91.68%.

Minimally invasive neurosurgery needs intraoperative imaging updates and high efficient image guide system to facilitate the procedure. An automatic image guided system utilized with a compact and mobile intraoperative CT imager was introduced in this work. A tracking frame that can be easily attached onto the commercially available skull clamp was designed. With known geometry of fiducial and tracking sensor arranged on this rigid frame that was fabricated through high precision 3D printing, not only was an accurate, fully automatic registration method developed in a simple and less-costly approach, but also it helped in estimating the errors from fiducial localization in image space through image processing, and in patient space through the calibration of tracking frame. Our phantom study shows the fiducial registration error as 0.348±0.028mm, comparing the manual registration error as 1.976±0.778mm. The system in this study provided a robust and accurate image-to-patient registration without interruption of routine surgical workflow and any user interactions involved through the neurosurgery.

Infrared medical examination finds the diseases through scanning the overall human body temperature and obtaining the temperature anomalies of the corresponding parts with the infrared thermal equipment. In order to obtain the temperature anomalies and disease parts, Infrared Medical Image Visualization and Anomalies Analysis Method is proposed in this paper. Firstly, visualize the original data into a single channel gray image: secondly, turn the normalized gray image into a pseudo color image; thirdly, a method of background segmentation is taken to filter out background noise; fourthly, cluster those special pixels with the breadth-first search algorithm; lastly, mark the regions of the temperature anomalies or disease parts. The test is shown that it’s an efficient and accurate way to intuitively analyze and diagnose body disease parts through the temperature anomalies.

B ultrasound as a kind of ultrasonic imaging, which has become the indispensable diagnosis method in clinical medicine. However, the presence of speckle noise in ultrasound image greatly reduces the image quality and interferes with the accuracy of the diagnosis. Therefore, how to construct a method which can eliminate the speckle noise effectively, and at the same time keep the image details effectively is the research target of the current ultrasonic image de-noising. This paper is intended to remove the inherent speckle noise of B ultrasound image. The novel algorithm proposed is based on both wavelet transformation of B ultrasound images and data fusion of B ultrasound images, with a smaller mean squared error (MSE) and greater signal to noise ratio (SNR) compared with other algorithms. The results of this study can effectively remove speckle noise from B ultrasound images, and can well preserved the details and edge information which will produce better visual effects.

Total variation(TV) based on regularization has been proven as a popular and effective model for image restoration, because of its ability of edge preserved. However, as the TV favors a piece-wise constant solution, the processing results in the flat regions of the image are easily produced "staircase effects", and the amplitude of the edges will be underestimated; the underlying cause of the problem is that the regularization parameter can not be changeable with spatial local information of image. In this paper, we propose a novel Scatter-matrix eigenvalues-based TV(SMETV) regularization with image blind restoration algorithm for deblurring medical images. The spatial information in different image regions is incorporated into regularization by using the edge indicator called difference eigenvalue to distinguish edges from flat areas. The proposed algorithm can effectively reduce the noise in flat regions as well as preserve the edge and detailed information. Moreover, it becomes more robust with the change of the regularization parameter. Extensive experiments demonstrate that the proposed approach produces results superior to most methods in both visual image quality and quantitative measures.

With the rapid development of medical imaging technology, medical image research and application has become a research hotspot. This paper offers a solution to non-rigid registration of medical images based on ordinal feature (OF) and manifold learning. The structural features of medical images are extracted by combining ordinal features with local linear embedding (LLE) to improve the precision and speed of the registration algorithm. A physical model based on manifold learning and optimization search is constructed according to the complicated characteristics of non-rigid registration. The experimental results demonstrate the robustness and applicability of the proposed registration scheme.

Segmentation of ultrasound image is an important port of ultrasound medical computer-aided system. However, due to the speckle noise, intensity heterogeneity, and low contrast, the ultrasonic segmentation is much difficult. In this paper, we introduce an effective method which integrates edge phase information and an effective active contour model to make the segmentation better. First, we use the speckle reducing anisotropic diffusion method to suppress the noise in ultrasound images. Then, we utilize the local phase information from monogenic signal to form a novel edge indicator and we use the indicator to replace the traditional intensity-based speed stopping term in distance regularized level set evolution. Another contribution of this paper is that we extend the proposed method to the field of ultrasonic thyroid nodules segmentation, qualitative and quantitative comparative results demonstrate the outperformance of our approach.

Images acquired during free breathing using contrast enhanced ultrasound hepatic perfusion imaging exhibits a periodic motion pattern. It needs to be compensated for if a further accurate quantification of the hepatic perfusion analysis is to be executed. To reduce the impact of respiratory motion, image-based breathing gating algorithm was used to compensate the respiratory motion in contrast enhanced ultrasound. The algorithm contains three steps of which respiratory kinetics extracted, image subsequences determined and image subsequences registered. The basic performance of the algorithm was to extract the respiratory kinetics of the ultrasound hepatic perfusion image sequences accurately. In this paper, we treated the kinetics extracted model as a non-negative matrix factorization (NMF) problem. We extracted the respiratory kinetics of the ultrasound hepatic perfusion image sequences by non-negative matrix factorization (NMF). The technique involves using the NMF objective function to accurately extract respiratory kinetics. It was tested on simulative phantom and used to analyze 6 liver CEUS hepatic perfusion image sequences. The experimental results show the effectiveness of our proposed method in quantitative and qualitative.

Brain tumor is one of the most fatal cancers, especially high-grade gliomas are among the most deadly. However, brain tumor MR images usually have the disadvantages of low resolution and contrast when compared with the optical images. Consequently, we present a novel adaptive intuitionistic fuzzy enhancement scheme by combining a nonlinear fuzzy filtering operation with fusion operators, for the enhancement of brain tumor MR images in this paper. The presented scheme consists of the following six steps: Firstly, the image is divided into several sub-images. Secondly, for each sub-image, object and background areas are separated by a simple threshold. Thirdly, respective intuitionistic fuzzy generators of object and background areas are constructed based on the modified restricted equivalence function. Fourthly, different suitable operations are performed on respective membership functions of object and background areas. Fifthly, the membership plane is inversely transformed into the image plane. Finally, an enhanced image is obtained through fusion operators. The comparison and evaluation of enhancement performance demonstrate that the presented scheme is helpful to determine the abnormal functional areas, guide the operation, judge the prognosis, and plan the radiotherapy by enhancing the fine detail of MR images.

It is well-known that the number of clusters is one of the most important parameters for automatic segmentation. However, it is difficult to define owing to the high diversity in appearance of tumor tissue among different patients and the ambiguous boundaries of lesions. In this study, a nonparametric mixture of Dirichlet process (MDP) model is applied to segment the tumor images, and the MDP segmentation can be performed without the initialization of the number of clusters. A new nonparametric segmentation algorithm combined with anisotropic diffusion and a Markov random field (MRF) smooth constraint is proposed in this study. Besides the segmentation of single modal brain tumor images, we developed the algorithm to segment multimodal brain tumor images by the magnetic resonance (MR) multimodal features and obtain the active tumor and edema in the same time. The proposed algorithm is evaluated and compared with other approaches. The accuracy and computation time of our algorithm demonstrates very impressive performance.

Calculation of ultrasonic field based on medical transducers is often done by applying acoustics and using the Tupholestetpanishen method of calculation. The calculation is based on spatial impulse response; the spatial impulse response has only been determined analytical for a few geometries and using apodization over the transducer surface generally make its impossible to find the response analytically. A popular approach to find the general field is thus to split the aperture into small rectangles, and then sum the weighted response from each of these. The problem with triangular is their poor fit apertures which do not have straight edges, such as circular and oval shapes. In order to solve the problem, a novel algorithm based on triangular be proposed in the paper, the simulation of ultrasonic field based on the algorithm can be improved obviously.

Nucleus and cytoplasm are both essential for white blood cell recognition but the edges of cytoplasm are too blurry to be detected because of instable staining and overexposure. This paper aims at proposing a cytoplasm enhancement operator (CEO) to achieve accurate convergence of the active contour model. The CEO contains two parts. First, a nonlinear over-exposure enhancer map is yielded to correct over-exposure, which suppresses background noise while preserving details and improving contrast. Second, the over-exposed regions of cytoplasm in particular is further enhanced by a tri- modal histogram specification based on the scale-space filtering. The experimental results show that the proposed CEO and its corresponding GVF snake is superior to other unsupervised segmentation approaches.

Lung respiratory movement can cause errors in the operation of image navigation surgery and they are the main errors in the navigation system. To solve this problem, the image-based motion correction strategy should be proposed to quickly correct the respiratory motion in the image sequence. So, the commercial ultrasound machine can display contrast and tissue images simultaneously. In the paper, a convenient, simple and easy-to-use breathing model whose precision was close to the sub-voxel was proposed. The first, in the clinical case the low gray-level variation in the tissue images, motion parameters were first calculated according to the actual lung movement information of each point the tissue images are registered by using template matching with sum of absolute differences metric. Finally, the similar images are selected by a double-selection method which requires global and local threshold setting. The generic breathing model was constructed based on all the sample data. The results of experiments show the algorithm can reduce the original errors caused by breath movement heavily.