Xuelong Hu

An Automatic Image Segmentation Algorithm Based on Weighting Fuzzy C-Means Clustering

Image segmentation is an important research topic in the field of computer vision. Now the fuzzy C-Means (FCM) algorithm is one of the most frequently used clustering algorithms. Although a FCM algorithm is a clustering without supervising, the FCM arithmetic should be given the transcendent information of prototype parameter; otherwise the arithmetic will be wrong. This limits its application in image segmentation. In this paper, we develop a new theoretical approach to automatically selecting the weighting exponent in the FCM to segment the image, which is called Automatic Clustering Weighting Fuzzy C-Means Segmentation (ACWFCM). This method can reduce the disturbance of noise; get the segmentation numbers more accurately. The experimental results illustrate the effectiveness of the proposed method.

A method for infrared image segment based on sharp frequency localized contourlet transform and morphology

A method associating sharp frequency localized contourlet transform with morphology to remove mixing noise and suppress background disturbance of infrared images is proposed. This method use sharp frequency localized contourlet transform and a morphology filter to deny the noise in infrared images. And then segment infrared images by regional method of morphology watershed, the target can be isolated easily. During the experiments, we can see that, the submit method can detect and segment the goal in the low SNR infrared image with complicated background effectively. And the result of this method is better than that based on contourlet transform and wavelet transform.

Flower Image Retrieval Based on Saliency Map

Most of general content-based image retrieval (CBIR) algorithms cannot meet the demand for fine retrieval of flower images. Combing with features of flower images, this paper proposed a flower image retrieval algorithm based on saliency map. Firstly, to obtain the saliency map, the improved Ittis visual attention model was utilized, and then the color and LBP texture feature were extracted using the saliency map, so as to the image segmentation was avoided. Finally, the retrieval experiments on flower image data sets of the VGG group were finished. Comparative results show that the proposed algorithm is more effective than the other two algorithms, i.e. color histogram combined with LBP texture histogram based on the original image (CT), and color and LBP texture histogram based on the saliency map extracted by Itti model (ICT).

Discriminative saliency propagation with sink points

Salient object detection is still very challenging especially in images with complex or cluttered background. In this paper, we present an efficient and discriminative framework to address it. In specially, a discriminative similarity metric is first proposed by measuring the chi-square distance in a new constructed feature space. Then, we apply it to calculate a background based coarse saliency map by introducing distribution prior to remove foreground noises in the image boundaries. Based on manifold ranking, a robust saliency propagation mechanism is further developed to highlight salient object and simultaneously suppress background region by setting appropriate sink points. Finally, several simple refinement techniques are utilized to generate pixel-wise and smooth saliency maps. Extensive experimental results show the superior performance of the proposed method in terms of different evaluation metrics. In addition, the proposed framework can be also applied to the existing saliency propagation methods for significant performance boosting. We also believe that it is a good choice for subsequent applications based on the achieved high performance and acceptable computational overhead. A discriminative similarity metric is proposed to distinguish similar regions.We propose an efficient distribution guided background based weak saliency method.A discriminative propagation mechanism is developed to optimize coarse saliency maps.Several new refinements are utilized to enhance the visual quality of the output.We make a good balance between saliency detection accuracy and computational cost.

Rotation-Invariant Texture Retrieval Based on Complementary Features

Among the traditional texture image retrieval algorithms based on wavelet transform, the limitation that the correlation between scales and sub bands is neglected leads to the poor retrieval efficiency. This paper proposed a novel rotation-invariant texture retrieval method which is based on complementary features. It firstly models the coefficients of sub bands with alpha-stable distribution and uses the fractional lower-order moment (FLOM) to capture the sub-Gaussian properties. Then estimate the so-called co variations between orientation sub bands as characteristic vectors of images. The next step is to construct a steer able multivariate sub-Gaussian model and deduce the rotation-invariant characteristic expression. Meanwhile make the low frequency energy statistics as part of the characteristics. Finally we choose a suitable distance function to measure the similarity between two images. The experimental results show that this method describes more image information, and it achieve a higher retrieval accuracy and it is a kind of effective way of rotation-invariant texture image retrieval.

Analysis of connectivity probability in VANETs considering minimum safety distance

Intelligent Transportation System (ITS), the safety-related message is propagated based on the wireless communications through vehicle to vehicle (V2V) and vehicle to infrastructures (V2I) in vehicular Ad hoc networks (VANETs) environments. Hence, the network connectivity is a key requirement for successfully safety-related message propagation. In this paper, we analyze the influence of vehicle density ρ and vehicles communication range R to the connectivity probability. Besides, a connectivity probability scheme which has considered the minimum safety distance Sd between adjacent vehicles is given. The analysis results indicate that small variation of the minimum safety distance will bring huge changes of the whole network connectivity. Therefore, the safety distance Sd cannot be ignored when designing the whole network connectivity models.

Enhancing V2V network connectivity for road safety by platoon-based VANETs

In vehicular ad-hoc networks(VANETs), road services related messages are propagated by vehicle to vehicle (V2V) and vehicle to infrastructure (V2I) communications. So the connectivity of VANETs is one key factor to ensure the successfully message dissemination. However, due to the dynamic changing topology of VANETs, the lifetime of the links between vehicles is short. Therefore, it is necessary to enhance the network connectivity for efficient message dissemination. In this paper, we propose a connectivity probability enhancing scheme by platoons, which also has considered the minimum safety distance between adjacent vehicles. The simulation results indicate the connectivity probability is always higher than those without platoons.

Approximate outage probability for multi-user vehicle cooperative communication

In order to achieve more efficient communication performances in vehicular ad hoc network (VANET) for road safety related message propagation, we propose a special nonbinary network coding scheme to improve the system reliability by reducing the outage probability of the communication system. Suppose all users in the network are provided with multiple antennas, which are responsible for message transmitting, relaying and receiving. Moreover, a cooperative phase is applied into the network. We analyze the exact expression of outage probability and give an approximation expression for calculating the outage probability over the Rayleigh fading channel. The simulation results indicate that the approximation expression has good performances on describing the exact expression.

Intelligent traffic light control system based on real time traffic flows

Currently, traffic congestions are the most serious issues that most cites are facing. In order to improving the urban traffic orders, as well as to alleviating traffic pressures, this paper presents an urban traffic control system, which is designed based on the real time traffic flow information. The proposed design has combined with traffic control theory, application of single chip computer and ultrasonic technology, design and research of the traffic control system based on traffic. Article control core of the system is the MCS − 51 single chip microcomputer, which achieves real-time monitoring by using ultrasonic sensors for road vehicle. Compared with the traditional control system, the system has the following characteristics: the duration time of traffic signal can be smartly set according to the number of road vehicles; a priority of lane can be assigned according to the actual demand when a vehicle is rarely at night, etc. Therefore, the traffic signals duration time can be smartly and intelligently adjusted according to the real time road traffic flow information.

A comparative study of object proposals re-ranking methods for object detection

Object detection is one of the most essential problems in computer vision and has made great progress in recent years, which is mainly contributed by powerful CNN and accurate object proposals. However, most of the existing proposal generation methods suffer from strong localization bias, and achieve high recall by outputting large amount bounding boxes (e.g. 2000). Thus, an effective proposal re-ranking approach becomes crucial to obtain high quality object proposals with few bounding box numbers. In this paper, we make a comparative study of the existing unsupervised objectness measuring approaches to testify their effectiveness and generalization abilities. Experiments on PASCAL VOC 2007 dataset demonstrates that contour is not an adequate objectness cue to pop out high quality proposals, because an object usually has a closed contour, while a good proposal (with high IoU) may not, and considering more objectness cues can get better performance. Thus, more effective objectness cues should be explored and combined together for proposals re-ranking, such more accurate saliency maps, which can greatly benefit subsequent object detection task and is also our future work.