Wenjun Tan

Dynamic hand gesture recognition using motion trajectories and key frames

In this paper, we address the problem of dynamic hand gesture recognition from unaided video sequences. We present a novel approach based on motion trajectories of hands and hand shapes of the key frames. Firstly, the hand area is segmented by active skin color model. Then the hand motion trajectories are extracted with dynamic time warping(DTW) algorithm and the key frame of video sequences is computed by frames difference. The hand gesture of the key frame is considered as a static hand gesture. The feature of hand shape is represented with Fourier descriptor and recognized by neural network. The combined method of the motion trajectories and key frame is presented to recognize the dynamic hand gesture from unaided video sequences. Experimental results show that the proposed approach is capable of effectively recognizing the dynamic hand gesture.

An Improved Algorithm of Hand Gesture Recognition under Intricate Background

This paper presents an integrated algorithm of YCbCr-Nrg, Double Color-Spatial Model and Background Model to resolve the problem that single skin-color model is obstructed by near kin color. This segmentation method is realized by the fusion of muti-feature. Based on the good describing ability of Fourier Descriptors algorithm and the good self-learning ability of BP neural network, an improved algorithm of hand recognition is presented and carried out. Results show that this algorithm is robustness for hand gesture recognition under intricate background.

Scene classification based on gray level-gradient co-occurrence matrix in the neighborhood of interest points

Scene classification is an important application field of multimedia information technology, whereas how to extract features from image is one of the key technologies in scene classification and recognition. A new method of extracting features is presented in this paper, it extracts features through gray level-gradient co-occurrence matrix in the neighborhood of interest points, also it can reserve the key image edge information, and it is called GGNP for short in the paper. The weighted Gowers similarity coefficient model is adopted as the basis for image scene classification, as it is more flexible than Euclidean distance function. Compared with traditional methods, the method has a good invariance in image scaling, rotation, translation and robust across a substantial range of affine distortion, meanwhile having good real-time. Experimentations are designed to test the precision and time-consuming of the method, the results of experiments show that the method has good effects on scene classification.

A novel interactive method of virtual reality system based on hand gesture recognition

Hand gesture is a natural and intuitive interactive method. This paper presents a novel interactive method of virtual reality system based on hand gesture recognition. The hand gesture segmentation method is proposed based on building complexion model by Gaussian distribution and the background model by automatically update the background parameters to improve the ability of adaptation environment. According to the good describing ability of Fourier Descriptor and the good self-learning ability of BP neural network, an improved algorithm of hand recognition is presented. Experiment result indicates that this method is flexible, realistic and exact, and fit for many virtual reality systems.

Hand extraction using geometric moments based on active skin color model

The hand gesture is the most common and natural way for human daily interaction. In this paper, we propose a novel approach to hand extraction based on active skin color model. The skin color model is first built by the non-linear transformation in YCbCr color space. Then the obtained model is applied to hand segmentation. Hand feature is extracted by calculating the seven moments of hand segmentation image. The experimental results demonstrate the feasibility of the proposed method.

Recognition of dynamic hand gesture based on SCHMM model

Dynamic gesture recognition is a key issue for visual gesture-based human-computer interaction. In this paper, a dynamic gesture recognition method is proposed based on SCHMM to solve the problem, which DHMM method is high speed and low rate and CHMM method is low speed and high rate. And the gesture clustering analysis, Baum—Welch algorithm of parametric estimation and Viterbi recognition algorithm is studied. Experiment result is given to show the balance performance of recognition rate and speed and the algorithm is satisfied for applications.

Fully automatic extraction of lung parenchyma from CT scans

In this paper, a novel fully automatic method of extraction of lung parenchyma is presented. Combining the iterative gray-level thresholds selection and the pulmonary regions extraction with error detection in 2D image, seed points and threshold are fast determined. In consequence, the pulmonary airspace is detected with 3D region growing method. Two steps airways segmentation with additional shape constrained criterion is used to completely remove airways from the airspace. To avoid lungs adhesion, the connections are detected and located. The dynamic programming method is applied to separate the left lung and right lung. Twelve clinical studies indicate that the novel method can meet the needs for quantification in diagnosis with respect to accuracy and time requirement.

A novel method for automated segmentation of airway tree

Accurate and automatic segmentation of airway tree from multi-slice computed tomography(MSCT) chest scan is an essential step for automatic computer aided diagnosing pulmonary diseases, including pulmonary emboli, pulmonary function and nodules detection. Due to the low contrast between airway walls and lung tissue on CT values, providing an accurate and in-vivo segmentation method for reconstructing of 3D anatomical tree structures from MSCT chest scan is a challenging issue for computer vision in medical imaging. In this paper a new-fully automated approach to segmentation airway tree is proposed. Firstly, the 3D seed point is extracted using the adaptive threshold algorithm in the first slice image, which the bronchi is demonstrated in the image. Secondly, the segmentation main bronchi with 3D region growing from the seed point and detection leaking into the lung parenchyma with computing the intergenerational volume are simultaneous computed. Thirdly, the probable leaking points are selected using simulating 3D region growing based on parallel computing when the leakage is detected. Finally, the segmentation bronchi is recycled with the selected seed points until up to the stopping condition. The results show that the proposed approach provides an automatic and efficient method to extract airway tree.

Research on Robotic Popular Science System Based on LEGO Bricks

The robotic technology is one of the hottest topic. With the increasingly wide range of applications in people¿s daily life, popularizing knowledge of robotic technology has become an important content of popular science. This paper researches autonomous navigation, visual identification, man-machine interaction and so on. According this, the robotic science educational systems with rectangular coordinates are designed using LEGO component. The experiment results show that the robotics¿ science educational system operated stably and reliably. It provides a novel approach to science education of robotics technology.

Research on the separation process model of pressure swing adsorption for CO 2 /N 2 with zeolite molecular sieve

The model of pressure swing adsorption separation process for CO2 with zeolite molecular sieve is established in this paper. The pressure swing adsorption system of CO2 is simulated with mass, gas and energy balance equations. The CO2 adsorption isotherm is fitted by Dual-site Langmuir equation. The recovery rate, productive rate and energy of the separation process is designed. Finally, the model is validated by CO2 of power plant flue gas with zeolite A and zeolite A+X. The result data of experiment and model simulation shows that our model is reliable in simulating the CO2/N2 pressure swing adsorption separation process with zeolite molecular sieve.