Search

Video object segmentation entails selecting and extracting objects of interest from a video sequence. Video Segmentation of Objects (VSO) is a critical task which has many applications, such as video edit, video decomposition and object recognition. The core of VSO system consists of two major problems of computer vision, namely object segmentation and object tracking. These two difficulties need to be solved in tandem in an efficient manner to handle variations in shape deformation, appearance alteration and background clutter. Along with segmentation efficiency computational expense is also a critical parameter for algorithm development. Most existing methods utilize advanced tracking algorithms such as mean shift and particle filter, applied together with object segmentation schemes like Level sets or graph methods. As video is a spatiotemporal data, it gives an extensive opportunity to focus on the regions of high spatiotemporal variation. We propose a new algorithm to concentrate on the high variations of the video data and use modified hierarchical processing to capture the spatiotemporal variation. The novelty of the research presented here is to utilize a fast object tracking algorithm conjoined with graph cut based segmentation in a hierarchical framework. This involves modifying both the object tracking algorithm and the graph cut segmentation algorithm to work in an optimized method in a local spatial region while also ensuring all relevant motion has been accounted for. Using an initial estimate of object and a hierarchical pyramid framework the proposed algorithm tracks and segments the object of interest in subsequent frames. Due to the modified hierarchal framework we can perform local processing of the video thereby enabling the proposed algorithm to target specific regions of the video where high spatiotemporal variations occur. Experiments performed with high frame rate video data shows the viability of the proposed approach.

Background and Purpose Complications due to brain edema and breakdown of blood brain barrier are an important factor affecting the treatment effects of patients with severe carotid stenosis. In this study, we investigated the protective effects of ischemic postconditioning on brain edema and disruption of blood brain barrier via establishing rat model of hypoperfusion due to severe carotid stenosis. Methods Wistar rat model of hypoperfusion due to severe carotid stenosis was established by binding a stainless microtube to both carotid arteries. Ischemic postconditioning procedure consisted of three cycles of 30 seconds ischemia and 30 seconds reperfusion. Brain edema was evaluated by measuring cerebral water content, and blood brain barrier permeability was assayed by examining cerebral concentration of Evans' Blue (EB) and fluorescein sodium (NaF). ELISA was used to analyze the expression of MMP-9, claudin-5 and occludin. The activity and location of MMP-9 was analyzed by gelatin zymography and in situ zymography, respectively. The distribution of tight junction proteins claudin-5 and occludin was observed by immunohistochemistry. Results The increased brain water content and cerebral concentration of EB and NaF were suppressed by administration of ischemic postconditioning prior to relief of carotid stenosis. Zymographic studies showed that MMP-9 was mainly located in the cortex and its activity was significantly improved by relief of carotid stenosis and, but the elevated MMP-9 activity was inhibited markedly by ischemic postconditioning. Immunohistochemistry revealed that ischemic postconditioning improved the discontinuous distribution of claudin-5 and occludin. ELISA detected that the expression of up-regulated MMP-9 and down-regulated claudin-5 and occludin caused by carotid relief were all attenuated by ischemic postconditioning. Conclusions Ischemic postconditioning is an effective method to prevent brain edema and improve BBB permeability and could be used during relief of severe carotid stenosis.

The effects of early relief of heavy bilateral carotid stenosis and ischemic postconditioning on hippocampus CA1 neurons are still unclear. In this study, we used a rat model to imitate severe bilateral carotid stenosis in humans. The rats were divided into sham group, carotid stenosis group, stenosis relief group and ischemic postconditioning group. Ischemic postconditioning consisted of three cycles of 30 s ischemia and 30 s reperfusion. The cerebral blood flow was measured with a laser Doppler flowmeter. Neuronal death in the CA1 region was observed by hematoxylin-eosin staining, and the number of live neurons was assessed by cell counting under a light microscope. The levels of oxidative products MDA and 8-iso-PGF2α, inflammatory factors IL-1β and TNF-α, and the activities of anti-oxidative enzymes SOD and CAT were assayed by specific enzyme-linked immunosorbent assay (ELISA) kits, respectively. We found that relief of carotid stenosis and ischemic postconditioning could increase cerebral blood flow. When stenosis was relieved, the percentage of live neurons was 66.6% ± 6.2% on day 3 and 62.3% ± 9.8% on day 27, which was significantly higher than 55.5% ± 4.8% in stenosis group. Ischemic postconditioning markedly improved the live neurons to 92.5% ± 6.7% on day 3 and 88.6% ± 9.1% on day 27. Further study showed that, neuronal death caused by relief of stenosis is associated with increased oxidative stress and enhanced inflammatory response, and the protection of ischemic postconditioning is related to inhibition of oxidative stress and suppression of inflammatory response.

This paper presents a novel method based on deterministic annealing to circumvent the problem of the sensitivity to atypical observations associated with the maximum likelihood (ML) estimator via conventional EM algorithm for mixture models. In order to learn the mixture models in a robust way, the parameters of mixture model are estimated by trimmed likelihood estimator (TLE), and the learning process is controlled by temperature based on the principle of maximum entropy. Moreover, we apply the proposed method to the single-trial electroencephalography (EEG) classification task. The motivation of this work is to eliminate the negative effects of artifacts in EEG data, which usually exist in real-life environments, and the experimental results demonstrate that the proposed method can successfully detect the outliers and therefore achieve more reliable result.

In this paper, we propose a modified visual stimulus generation method and feature detection algorithm to design a frequency and phase coding steady-state visual evoked potential (SSVEP) based brain-computer interface (BCI). By utilizing both frequency and phase information, we solve the harmonic frequency problem in our proposed SSVEP-BCI system. The offline experimental results show that the proposed feature detection algorithm can enhance the classification rate over 10% (from 69%±12% to 82%±8%) even though only one signal electrode is used and the harmonic frequencies (6.67Hz, 13.33Hz, 8.57Hz and 17.14Hz) are employed.

In this paper, we propose an entropy penalized approach to address the problem of learning the parameters of Gaussian mixture models (GMMs) with components of small weights. In addition, since the method is based on minimum message length (MML) criterion, it can also determine the number of components of the mixture model. The simulation results demonstrate that our method outperform several other state-of-art model selection algorithms especially for the mixtures with components of very different weights.

This paper presents a steady-state visual evoked potential (SSVEP) based online brain-computer interface (BCI) chatting system. In this application, it helps disable subjects to chat by means of their brain signal activities. Stimuli illuminated at different frequencies are displayed on liquid crystal display (LCD) screen. The BCI system implements a chatting interface programmed in Java, and it provides a convenient and high efficient communication platform for paralyzed patients. A novel idle state detection is introduced in this proposed system, and a high transfer rate of 63.54 bits/min and 92% idle state detection accuracy are achieved.

This paper presents an online steady-state visual evoked potential (SSVEP) based brain-computer interface (BCI). Stimuli are displayed on a liquid crystal display (LCD) screen with a frame based encoding method to elicit SSVEPs with a wide range of frequencies. This system focuses strongly on practicability and convenience, including an adequate alphabet (42 characters) that can allow a wide range of options. Four healthy subjects attain a mean information transfer rate (ITR) of 61.64±3.61 bits/min.

Minimum energy combination (MEC) and canonical correlation analysis (CCA) are widely used for steady-state visual evoked potential (SSVEP) based brain computer interface (BCI), since both approaches have satisfactory performance. The purpose of this paper is to provide a guideline on choice of detection method, through comparison of the performance of the two approaches from simulation data and real SSVEP data. The experiment results show that CCA has lower deviation, higher accuracy and higher signal to noise ratio than MEC.

The performance of a brain computer interface (BCI) system is usually degraded due to the outliers in electroencephalography (EEG) samples. This paper presents a novel outlier detection method based on robust learning of Gaussian mixture models (GMMs). We apply the proposed method to the single-trial EEG classification task. After trial-pruning, feature extraction and classification are performed on the subset of training data, and experimental results demonstrate that the proposed method can successfully detect the outliers and therefore achieve more reliable result.

Gaussian mixture model (GMM) has been considered to model the EEG data for the classification task in brain-computer interface (BCI) system. In the practical BCI application, however, the performance of the classical GMM optimized by standard expectation-maximization (EM) algorithm may be degraded due to the noise and outliers, which often exist in realistic BCI systems. The motivation of this paper is to introduce the GMM based on the combination between the genetic algorithm (GA) and EM method to give a probabilistic output for further analysis and, more important, to achieve the reliable estimation by pruning the potential outliers and noisy samples in the EEG data, so the performance of BCI system can be improved. Experiments on two BCI datasets demonstrate the improvement in comparison with the classical mixture model.

This paper proposes an improved method for generating different phases of visual stimulus while liquid crystal display (LCD)/cathode ray tube (CRT) is employed as the visual stimulator. Since using the traditional method can only generate the limited frequencies and phases of visual stimulus, increasing the number of different flickering targets becomes very difficult in steady-state visual evoked potential (SSVEP)-based brain-computer interface (BCI). The experimental result shows that the proposed method can generate the visual stimulus with more phase angles than the traditional method. In addition, the proposed LCD visual stimulator can evoke the subject's SSVEP with expected phase.

Due to the artifacts in electroencephalography (EEG) data, the performance of brain-computer interface (BCI) is degraded. On the other hand, in the motor imagery based BCI system, EEG signals are usually contaminated by the misleading trials caused by improper imagination of a movement. In this paper, we present a novel algorithm to detect the abnormal EEG data using genetic algorithm (GA). After trial pruning, a subset of the EEG data are selected, on which common spatial pattern (CSP) and Gaussian classifier are trained. The performance of the proposed method is tested on Data set IIa of BCI Competition IV, and the simulation result demonstrates a significant improvement for six out of nine subjects.

The performances of different off-line methods for two different Electroencephalograph (EEG) signal classification tasks - motor imagery and finger movement, are investigated in this paper. The classifiers based on linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), kernel fisher discriminant (KFD), support vector machine (SVM), multilayer perceptron (MLP), learning vector quantization (LVQ) neural network, k-nearest neighbor (k-NN), and decision tree (DT), are compared in terms of classification accuracy. The main purpose of this paper is to provide a fair and extensive comparison of some commonly employed classification methods under the same conditions so that the assessment of different classifiers will be more convictive. As a result, a guideline for choosing appropriate algorithms for EEG classification tasks is provided.

Classification of electroencephalogram (EEG) is an important and challenging issue for brain computer interface (BCI) system. In this paper, an algorithm based on common spatial subspace decomposition (CSSD) and support vector clustering (SVC) is proposed to classify single-trial EEG recording during left or right finger movement. The algorithm is tested by the dataset IV of "BCI competition 2003", and the experimental result shows the proposed method, only using bereitschaftspotential (BP), rather than both BP and event-related desynchronization (ERD), has higher classification accuracy than the best one reported in the competition.

The clinical need for organ- or tissue-specific drug delivery, also known as targeted drug delivery, arises when systemic delivery of a drug in sufficient doses to achieve a therapeutic effect at the target site results in deleterious systemic effects. Relevant clinical problems include delivery of chemotherapeutic drugs to tumors, delivery of thrombolytic drugs to the cerebral or coronary circulation during ischemic stroke or myocardial infarction, and delivery of vasoactive drugs to the cerebral or coronary circulation for treatment of vasospasm. Current solutions to these problems involve invasive catheterization procedures directed at selective intra-arterial infusion of a drug into the target vascular bed. Such procedures often carry major risks. In many cases, therapeutic efficacy is lost shortly after removal of the indwelling catheter and cessation of drug infusion. Additional treatments require repetition of the invasive catheterization procedures with incremental accumulation of procedure- related risks. In most cases, prolonged maintenance of indwelling catheters for sustained or repetitive treatment is associated with prohibitive risk. Over the last 4 decades, liposomes have been explored as targeted drug delivery vehicles.1,2 A number of clinical and experimental therapeutic agents have been successfully loaded into liposomes.3 Liposomes, which are formed by enclosure of an aqueous core by 1 or more phospholipid bilayers, range in size from about 100 nm to several micrometers. Liposomes offer several advantages as drug delivery vehicles. Water-soluble compounds may be entrapped in the aqueous core, whereas water-insoluble compounds may bind to the lipid membrane. A liposome-associated drug (either bound to the lipid or entrapped in the aqueous core) is hypothesized to remain physiologically inactive until the physical integrity of the lipid membrane is disturbed. Drug-loaded liposomes propagated in the circulation can function as mobile drug reservoirs with the ability to deliver large pay-loads of a drug to specific target organs or tissues when appropriately triggered. Several chemical and physical methods for triggered drug release from liposomes have been explored, including acoustic pressure-, temperature-, and pH-dependent mechanisms.4–6 One promising approach that is being explored for drug delivery is the use of ultrasound-activated echogenic liposomes (ELIPs). These agents are synthesized by incorporating air and a drug of interest into liposomes that are submicrometer in size (mean diameter, 780 nm).7 These encapsulated air pockets are probably on the order of tens or hundreds of nanometers in diameter and thus can be considered nanobubbles. Note that the perfluorocarbon gas pockets in many contrast agents used clinically today are on the order of several micrometers and thus encapsulate microbubbles, not nanobubbles.8,9 Exposure of ELIPs to suitable pulses of ultrasound disturbs the physical structure of the liposome and results in drug release. Because ELIP activation by ultrasound can be controlled both spatially and temporally, ELIPs are potentially powerful tools for selective tissue- or organ-specific drug delivery. Furthermore, the encapsulated nanobubbles can serve as ultrasonic contrast agents, enabling the process of drug release and activation to be imaged in real time. The pulse repetition frequency and duty cycle dependence of the acoustic pressure destruction threshold for ELIPs has been characterized with 6-MHz pulsed Doppler ultrasound.10 Continuous wave ultrasound (2–7 W/cm2) has been investigated as a means of triggering calcein or recombinant tissue plasminogen activator (rt-PA) release from ELIP in nonflowing solutions.4,11,12 Recently, pulsed color Doppler ultrasound was used to trigger release of rt-PA from rt-PA-loaded ELIPs in flowing solutions.13 The use of color Doppler ultrasound (a scanned mode) enables a larger number of ELIPs to be exposed per unit of time than spectral Doppler ultrasound (an unscanned mode). This is important in clinical drug delivery paradigms because ELIPs circulating in the bloodstream will move rapidly through the ultrasound field. Because in vitro evaluation of ultrasound-triggered drug release must consider this aspect of clinical drug delivery, our experimental design used a flow model with a peristaltic pump to circulate ELIP solutions through an ultrasound exposure volume. Calcein, a water-soluble polyanionic fluorescein derivative with photosensitizing properties, is used as a fluorescent indicator in drug release studies as well as a contrast agent in retinal angiography.14,15 Papaverine, a spasmolytic lipophilic opium alkaloid, is administered into cerebral arteries through a microcatheter for clinical treatment of posthemorrhagic cerebral vasospasm.16,17 In this study, calcein-loaded ELIPs (C-ELIPs) and papaverine-loaded ELIPs (P-ELIPs) were circulated in a flow model and exposed to pulsed color Doppler ultrasound to trigger calcein or papaverine release from C-ELIPs or P-ELIPs. The objective of this study was to quantify the release of calcein and papaverine from C-ELIPs and P-ELIPs, respectively. Dynamic changes in echogenicity were also assessed with B-mode ultrasound before, during, and after color Doppler ultrasound exposure.

In a control chart, unnatural patterns are always associated with some specific assignable causes that should be eliminated. The identification of control chart pattern (CCP) is therefore important and further estimation of the unnatural pattern parameters can improve the manufacturing process. In this paper, a modified counter-propagation network (m-CPN) is developed to classify the mean shift and simultaneously estimate the shift magnitude. The m-CPN is compared with five existing networks through numerical simulation and the result shows a better performance of the m-CPN in terms of classification accuracy, as well as both Type I and Type II errors.

This investigation elucidates the Akt/mitochondrial ATP-sensitive K+ (mitoKATP) channel signaling pathway in late pharmacological preconditioning, using the mitoKATP channel openers BMS-191095 (BMS) and diazoxide (DE). BMS (1 mg/kg ip) and DE (7 mg/kg ip) alone or BMS plus wortmannin (WTN, 15 μg/kg ip), an inhibitor of phosphatidylinositol 3-kinase, and BMS plus 5-hydroxydecanoic acid (5-HD, 5 mg/kg ip), an inhibitor of mitoKATP channels, were administered to male mice. Twenty-four hours later, hearts were isolated and subjected to 40 min of ischemia and 120 min of reperfusion via Langendorff's apparatus. Both BMS and DE reduced left ventricular end-diastolic pressure and increased left ventricular developed pressure as well as reduced LDH release. Coadministration of BMS and WTN abolished the beneficial effects of BMS on cardiac function. Moreover, BMS and DE accelerated Akt phosphorylation in cardiac tissue as determined by Western blot analysis and also significantly reduced apoptosis compared with ischemic control. WTN significantly suppressed BMS-induced Akt phosphorylation, whereas 5-HD had no effect on Akt phosphorylation in cytosol, and the effect of BMS on apoptosis was abolished. It is concluded that the cardioprotective effect by mitoKATP channels is attributed to the translocation of phosphorylated Akt from cytosol to mitochondria.

A deterministic annealing (DA)-based expectation-maximisation (EM) algorithm is proposed for robust learning of Gaussian mixture models. By combing the DA approach, trimmed likelihood function and Bayesian information criterion (BIC), the proposed algorithm can simultaneously perform model selection and outlier detection, and mitigate the problems of local optima and boundary of parameter space with the conventional EM algorithm. Experiments demonstrate that the proposed algorithm can determine the number of components correctly even though the data are contaminated by outliers.

We present a novel multitask learning framework called multitask generalized eigenvalue program (MTGEP), which jointly solves multiple related generalized eigenvalue problems (GEPs). This framework is quite general and can be applied to many eigenvalue problems in machine learning and pattern recognition, ranging from supervised learning to unsupervised learning, such as principal component analysis (PCA), Fisher discriminant analysis (FDA), common spatial pattern (CSP), and so on. The core assumption of our approach is that the leading eigenvectors of related GEPs lie in some subspace that can be approximated by a sparse linear combination of basis vectors. As a result, these GEPs can be jointly solved by a sparse coding approach. Empirical evaluation with both synthetic and benchmark real world datasets validates the efficacy and efficiency of the proposed techniques, especially for grouped multitask GEPs.