Yajun Fang

Comparison between infrared-image-based and visible-image-based approaches for pedestrian detection

In order to improve the safety of night driving, automatic pedestrian detection has received more and more attraction. Since reliability is the most important issue in these systems, multi-dimensional-feature-based segmentation and classification needs to be introduced, and each axis should be efficient and be as much independent (to each other) as possible. To choose effective multi-dimensional features for infrared-image-based detection, the paper first investigates the possibilities of reusing available features for visible images by analyzing the different properties of infrared images and visible images. To take advantage of unique properties of infrared images, we propose the following novel features: special projection feature for segmentation, and two-axis pixel-distribution feature for classification. The segmentation based on new features does not depend on many assumptions and is shape-independent, thus avoiding brute-force multiple templates and multi-scale pyramid searching. The novel classification features include histogram feature and inertial feature that are independent and complimentary, thus the two-dimensional fusion-based classification significantly improves detection accuracy. These proposed features are independent from conventional pixel-array feature, and can be further fused with other general pedestrian detection features to improve simplicity, speed, and reliability.

Depth-based target segmentation for intelligent vehicles: fusion of radar and binocular stereo

Dynamic environment interpretation is of special interest for intelligent vehicle systems. It is expected to provide lane information, target depth, and the image positions of targets within given depth ranges. Typical segmentation algorithms cannot solve the problems satisfactorily, especially under the high-speed requirements of a real-time environment. Furthermore, the variation of image positions and sizes of targets creates difficulties for tracking. In this paper, we propose a sensor-fusion method that can make use of coarse target depth information to segment target locations in video images. Coarse depth ranges can be provided by radar systems or by a vision-based algorithm introduced in the paper. The new segmentation method offers more accuracy and robustness while decreasing the computational load.

Distance/motion-based segmentation under heavy background noise

Typical segmentation algorithms are challenged by background noise and the variation of object sizes and object positions in video frames. In this paper, we propose a new object segmentation method based on both motion and distance information to increase segmentation reliability and to suppress background noise. Two new concepts are described in this paper. First proposed is a new distance-based background detection algorithm to remove the impact of noisy background without using reference frames. The second proposed is a new depth/motion-based segmentation that can accurately capture objects of different sizes. The algorithm introduced successfully increases the accuracy and reliability of object segmentation and motion detection.

TV camera-based vehicle motion detection and its chip implementation

Detecting motion of vehicles and other objects by using TV cameras has wide applications in video image compression (MPEG 2) and intelligent transportation systems. In this paper, we present an edge-based motion detection to overcome the heavy computational load of the conventional motion detection algorithm. The algorithm successfully decreases both the computational load and the area of chip implementation by factors of 4.4 and 8, respectively. The result from the chip illustrates the effectiveness of the algorithm.

A layered-based fusion-based approach to detect and track the movements of pedestrians through partially occluded situations

To obtain perception abilities, conventional methods independently detect static and dynamic obstacles, and estimate their related information, which is not quite reliable and computationally heavy. We propose a fusion-based and layered-based approach to systematically detect dynamic obstacles and obtain their location and timing information. The layered-based concept helps us to first search pedestrians in horizontal dimension based on transitional peaks in the defined projection-curves, and then search in vertical dimension. Converting a typical 2D search problem into two 1D search problems significantly decreases the computational load. The fusion-based obstacle detection fuses the information from initial segmentation and dynamic tracking model to avoid complicated tracking schemes. The methodologies take advantage of connection between different information, and increase the accuracy and reliability of obstacle segmentation and tracking. The search mechanism works for both visible and infrared sequences, and is specifically effective to track the movements of pedestrians in complicated environments such as human intersecting and conclusion, thus improving environment understanding abilities and driving safety.

A Clustering Validity Index Based on Pairing Frequency

Clustering is an important problem, which has been applied in many research areas. However, there is a large variety of clustering algorithms and each could produce quite different results depending on the choice of algorithm and input parameters, so how to evaluate clustering quality and find out the optimal clustering algorithm is important. Various clustering validity indices are proposed under this background. Traditional clustering validity indices can be divided into two categories: internal and external. The former is mostly based on compactness and separation of data points, which is measured by the distance between clusters’ centroids, ignoring the shape and density of clusters. The latter needs external information, which is unavailable in most cases. In this paper, we propose a new clustering validity index for both fuzzy and hard clustering algorithms. Our new index uses pairwise pattern information from a certain number of interrelated clustering results, which focus more on logical reasoning than geometrical features. The proposed index overcomes some shortcomings of traditional indices. Experiments show that the proposed index performs better compared with traditional indices on the artificial and real datasets. Furthermore, we applied the proposed method to solve two existing problems in telecommunication fields. One is to cluster serving GPRS support nodes in the city Chongqing based on service characteristics, the other is to analyze users’ preference.

Heterogeneous characters modeling of instant message services users’ online behavior

Research on temporal characteristics of human dynamics has attracted much attentions for its contribution to various areas such as communication, medical treatment, finance, etc. Existing studies show that the time intervals between two consecutive events present different non-Poisson characteristics, such as power-law, Pareto, bimodal distribution of power-law, exponential distribution, piecewise power-law, et al. With the occurrences of new services, new types of distributions may arise. In this paper, we study the distributions of the time intervals between two consecutive visits to QQ and WeChat service, the top two popular instant messaging services in China, and present a new finding that when the value of statistical unit T is set to 0.001s, the inter-event time distribution follows a piecewise distribution of exponential and power-law, indicating the heterogeneous character of IM services users’ online behavior in different time scales. We infer that the heterogeneous character is related to the communication mechanism of IM and the habits of users. Then we develop a combination model of exponential model and interest model to characterize the heterogeneity. Furthermore, we find that the exponent of the inter-event time distribution of the same service is different in two cities, which is correlated with the popularity of the services. Our research is useful for the application of information diffusion, prediction of economic development of cities, and so on.