38 research outputs found
Hierarchical background subtraction using local pixel clustering
No full textWe propose a robust hierarchical background subtraction technique which takes the spatial relations of neighboring pixels in a local region into account to detect objects in difficult conditions. Our algorithm combines a per-pixel with a per-region background model in a hierarchical manner, which accentuates the advantages of each. This is a natural combination because the two models have complementary strengths. The per-pixel background model is achieved by mixture of Gaussians Models (GMM) with RGB feature. Although precisely describing background change in high resolution, it suffers from the sensitivity to quick variations in dynamic environment. To tolerate these quick variations, we further develop a novel GMM based per-region background model, which is updated by the cluster centers obtained from a k-means clustering of the pixels' RGB feature in the region. Numerical and qualitative experimental results on challenging videos demonstrate the robustness of the proposed method. ? 2008 IEEE.EI
Jointly Feature Learning and Selection for Robust Tracking via a Gating Mechanism.
No full textTo achieve effective visual tracking, a robust feature representation composed of two separate components (i.e., feature learning and selection) for an object is one of the key issues. Typically, a common assumption used in visual tracking is that the raw video sequences are clear, while real-world data is with significant noise and irrelevant patterns. Consequently, the learned features may be not all relevant and noisy. To address this problem, we propose a novel visual tracking method via a point-wise gated convolutional deep network (CPGDN) that jointly performs the feature learning and feature selection in a unified framework. The proposed method performs dynamic feature selection on raw features through a gating mechanism. Therefore, the proposed method can adaptively focus on the task-relevant patterns (i.e., a target object), while ignoring the task-irrelevant patterns (i.e., the surrounding background of a target object). Specifically, inspired by transfer learning, we firstly pre-train an object appearance model offline to learn generic image features and then transfer rich feature hierarchies from an offline pre-trained CPGDN into online tracking. In online tracking, the pre-trained CPGDN model is fine-tuned to adapt to the tracking specific objects. Finally, to alleviate the tracker drifting problem, inspired by an observation that a visual target should be an object rather than not, we combine an edge box-based object proposal method to further improve the tracking accuracy. Extensive evaluation on the widely used CVPR2013 tracking benchmark validates the robustness and effectiveness of the proposed method
Poliumoside inhibits apoptosis, oxidative stress and neuro-inflammation to prevent intracerebroventricular streptozotocin-induced cognitive dysfunction in Sprague–Dawley rats: an in vivo, in vitro and in silico study
Get PDFBACKGROUND: Alzheimer’s disease (AD) is a severe neurological illness that causes cognitive decline and death if not treated early. The current therapeutic modalities are inefficient in managing the cognitive dysfunction of AD. Therefore, in this study, we aimed to investigate the pharmacological benefit of poliumoside (PMD) in streptozotocin-induced cognitive dysfunction in Sprague–Dawley (SD) rats. MATERIALS AND METHODS: Initially, cognitive dysfunction in rats was induced by the intracerebroventricular administration of streptozotocin. Then rats received PMD at 5 mg and 10 mg/kg body weight. Various behavioural analyses, such as the Morris water maze (MWM) and the object recognition test (ORT), and locomotor analysis was conducted in the PMD-treated group. Biochemical analysis was conducted to analyse the effect of PMD on hippocampus oxidative-nitrosative stress and pro-inflammatory cytokines. MTT assay and annexin V/PI staining were performed to analyse the effect of PMD on the cell viability and neuronal toxicity of PC12 cells, respectively. Molecular docking analysis was also conducted with crystal structure of human AChE. RESULTS: PMD treatment improved cognitive capacity in rats in MWM and ORT. Compared to STZ rats, PMD-treated rats had significantly higher locomotor activity and lower AChE activity. PMD also restored dopamine, 5-HT, and NE levels and reduced their metabolic deactivation, as evidenced by increased levels of DOPAC, HVA, and 5-HIAA. Nitrite, MDA, SOD, CAT, and GSH levels were restored to near normal in PMD-treated rats, reducing hippocampus oxidative-nitrosative stress. Pro-inflammatory cytokines were similarly lowered in PMD-treated rats. In in vitro studies, PMD did not affect PC12 cell survival at the maximal dose of 10 μM. In addition, PMD concentration-dependently prevented H₂O₂-induced neuronal death in PC12 cells. In silico docking analysis showed that PMD fitted snugly into the active site of human AChE by engaging with the anionic domain and the catalytic triad of Trp86, Tyr337, Phe338, and Gly121 residues. CONCLUSIONS: This study has demonstrated that PMD has a significant impact on AD by inhibiting ACheE and restoring neurotransmitter levels, which enhances Ach levels in rats, and improves cognitive impairment in STZ rats
Medical Image Segmentation Based on Spatially Constrained Inverted Beta-Liouville Mixture Models
No full textSigma Set Based Implicit Online Learning for Object Tracking
No full textThis letter presents a novel object tracking approach within the Bayesian inference framework through implicit online learning. In our approach, the target is represented by multiple patches, each of which is encoded by a powerful and efficient region descriptor called Sigma set. To model each target patch, we propose to utilize the online one-class support vector machine algorithm, named Implicit online Learning with Kernels Model (ILKM). ILKM is simple, efficient, and capable of learning a robust online target predictor in the presence of appearance changes. Responses of ILKMs related to multiple target patches are fused by an arbitrator with an inference of possible partial occlusions, to make the decision and trigger the model update. Experimental results demonstrate that the proposed tracking approach is effective and efficient in ever-changing and cluttered scenes.Engineering, Electrical & ElectronicSCI(E)EI0ARTICLE9807-8101
Extracting Target Detection Knowledge Based on Spatiotemporal Information in Wireless Sensor Networks
No full textWireless sensor networks (WSNs) have been deployed for many applications of target detection, such as intrusion detection and wildlife protection. In these applications, the first step is to detect whether the target is present or not. However, most of the existing work uses the “simple disk model” as signal model, which may not capture the sensing environment. In this work, we utilize a more realistic signal model to describe sensing process of sensors. On the other hand, the “majority rule” is widely used to make the final decision, which may not obtain the true judgment. To this end, we utilize a more realistic signal model and also use a probabilistic decision model to make the final decision. Moreover, we propose a probabilistic detection algorithm in which all sensors' local measurement values are fully used. This algorithm does not need any artificial threshold compared with traditional algorithms. It makes the most of spatiotemporal information to obtain the final decision. For the spatial perspective, sensors are distributed in different locations cooperating with each other. Meanwhile, for the temporal perspective, multiround subdecisions are fused. The effectiveness of the proposed method is validated by extensive simulation results, which show high detection probabilities and low false alarm probabilities
Kernel correlation filters for visual tracking with adaptive fusion of heterogeneous cues
No full text
