养目八法

养目八法文／林涛图／大忠熨眼法两手掌相互摩擦发热后，按抚于双眼上。反复３次后，再以食指、中指、无名指轻压眼球。此法可改善限部血液循环，使双眼变得明亮。运目法即运转眼球，上下左右反复转动。交替平视、远视、近视、凝视物体，以锻炼眼球，改善视力。薰浴法以热...

An improved method for multiple targets tracking

The difficulty of multiple targets tracking is how to quickly fulfill the target matching from one flame image to another and fix the position of the target. In order to accurately choose target feature information for reliable matching, simplify operations under the reliable precondition, and realize precise moving objects tracking, an approach based on Kalman prediction and feature matching was proposed. The position of the target in next frame image was predicted by Kalman, and then the moving objects of two adjacent frames were matched by the centroid and area methods. When occlusion occurs, the best matching result was found to realize tracking by matching matrix algorithm. The simulation results show that the proposed method can achieve multiple targets tracking accurately and in real-time under complicated motion movements.

An improved Vibe algorithm for illumination mutations

The visual background extractor(Vibe)algorithm can lead to a large area of false detection in the extracted foreground target when the illumination is mutated.An improved Vibe method based on the YCbCr color space and improved three-frame difference is proposed in this paper.The algorithm detects the illumination mutation frames accurately based on the difference between the luminance components of two frames adjacent to a video frame.If there exists a foreground moving target in the previous frame of the mutated frame,three-frame difference method is utilized;otherwise,Vibe method using current frame is used to initialize background.Improved three-frame differential method based on the difference in brightness between two frames of the video changes the size of the threshold adaptively to reduce the interference of noise on the foreground extraction.Experiment results show that the improved Vibe algorithm can not only suppress the“ghost”phenomenon effectively but also improve the accuracy and completeness of target detection,as well as reduce error rate of detection when the illumination is mutated.

Automatic target recognition of moving target based on empirical mode decomposition and genetic algorithm support vector machine

In order to improve measurement accuracy of moving target signals, an automatic target recognition model of moving target signals was established based on empirical mode decomposition(EMD) and support vector machine(SVM). Automatic target recognition process on the nonlinear and non-stationary of Doppler signals of military target by using automatic target recognition model can be expressed as follows. Firstly, the nonlinearity and non-stationary of Doppler signals were decomposed into a set of intrinsic mode functions(IMFs) using EMD. After the Hilbert transform of IMF, the energy ratio of each IMF to the total IMFs can be extracted as the features of military target. Then, the SVM was trained through using the energy ratio to classify the military targets, and genetic algorithm(GA) was used to optimize SVM parameters in the solution space. The experimental results show that this algorithm can achieve the recognition accuracies of 86.15%, 87.93%, and 82.28% for tank, vehicle and soldier, respectively.

A combination algorithm of Chaos optimization and genetic algorithm and its application in maneuvering multiple targets data association

The most important problem in targets tracking is data association which may be represented as a sort of constraint combinational optimization problem. Chaos optimization and adaptive genetic algorithm were used to deal with the problem of multi-targets data association separately. Based on the analysis of the limitation of chaos optimization and genetic algorithm, a new chaos genetic optimization combination algorithm was presented. This new algorithm first applied the "rough" search of chaos optimization to initialize the population of GA, then optimized the population by real-coded adaptive GA. In this way, GA can not only jump out of the "trap" of local optimal results easily but also increase the rate of convergence. And the new method can also avoid the complexity and time-consumed limitation of conventional way. The simulation results show that the combination algorithm can obtain higher correct association percent and the effect of association is obviously superior to chaos optimization or genetic algorithm separately. This method has better convergence property as well as time property than the conventional ones.

Area Variation Based Color Snake Algorithm for Moving Object Tracking

A snake algorithm has been known that it has a strong point in extracting the exact contour of an object. But it is apt to be influenced by scattered edges around the control points. Since the shape of a moving object in 2D image changes a lot due to its rotation and translation in the 3D space, the conventional algorithm that takes into account slowly moving objects cannot provide an appropriate solution. To utilize the advantages of the snake algorithm while minimizing the drawbacks, this paper proposes the area variation based color snake algorithm for moving object tracking. The proposed algorithm includes a new energy term which is used for preserving the shape of an object between two consecutive images. The proposed one can also segment precisely interesting objects on complex image since it is based on color information. Experiment results show that the proposed algorithm is very effective in various environments.

A robust system for real-time pedestrian detection and tracking

A real-time pedestrian detection and tracking system using a single video camera was developed to monitor pedestrians. This system contained six modules: video flow capture, pre-processing, movement detection, shadow removal, tracking, and object classification. The Gaussian mixture model was utilized to extract the moving object from an image sequence segmented by the mean-shift technique in the pre-processing module. Shadow removal was used to alleviate the negative impact of the shadow to the detected objects. A model-free method was adopted to identify pedestrians. The maximum and minimum integration methods were developed to integrate multiple cues into the mean-shift algorithm and the initial tracking iteration with the competent integrated probability distribution map for object tracking. A simple but effective algorithm was proposed to handle full occlusion cases. The system was tested using real traffic videos from different sites. The results of the test confirm that the system is reliable and has an overall accuracy of over 85%.

Moving Target Tracking Based on LogGabor Wavelet and Mean-shift Algorithm

Tracking moving target from image or video sequence is a hot research topic in computer vision. An algorithm based on LogGabor wavelet and Mean-shift has been proposed for moving target tracking under fixed camera setting and complicated environment. Phase coherency of LogGabor wavelet facilitates to extract the edge of moving target and check noise. According to the edge detection,the starting location of Mean-shift can be estimated using the target center coordinate. Eventually,a real-time moving target can be extracted by doing iterative matching pursuit,and experimental results proved the effectiveness of the method proposed.

Finger Flexion Motion Inference from sEMG Signals

This paper provides a method to infer finger flexing motions using a 4-channel surface Electronyogram （sEMG）. Surface EMGs are hannless to the humnan body and easily done. However, they do not reflect the activity of specific nerves or muscles, unlike invasive EMCs. On the other hand, the non-invasive type is difficult to use for discriminating various motions while using only a small number of electrodes. Surface EMG data in this study were obtained from four electodes placed around the forearm. The motions were the flexion of each 5 single fingers （thumb, index finger, middle finger, ring finger, and little fingers）. One subject was trained with these motions and another left was untrained. The maximum likelihood estimation method was used to infer the finger motion. Experimental results have showed that this method could be useful for recognizing finger motions.The average accuracy was as high as 95%.