Sin-Yu Chen

Carried Object Detection Using Ratio Histogram and its Application to Suspicious Event Analysis

This letter proposes a novel method to detect carried objects from videos and applies it for analysis of suspicious events. First of all, we propose a novel kernel-based tracking method for tracking each foreground object and further obtaining its trajectory. With the trajectory, a novel ratio histogram is then proposed for analyzing the interactions between the carried object and its owner. After color re-projection, different carried objects can be then accurately segmented from the background by taking advantages of Gaussian mixture models. After bag detection, an event analyzer is then designed to analyze various suspicious events from the videos. Even though there is no prior knowledge about the bag (such as shape or color), our proposed method still performs well to detect these suspicious events. As we know, due to the uncertainties of the shape and color of the bag, there is no automatic system that can analyze various suspicious events involving bags (such as robbery) without using any manual effort. However, by taking advantages of our proposed ratio histogram, different carried bags can be well segmented from videos and applied for event analysis. Experimental results have proved that the proposed method is robust, accurate, and powerful in carried object detection and suspicious event analysis.

Vehicle Color Classification Under Different Lighting Conditions Through Color Correction

This paper presents a novel color correction technique for classifying vehicles under different lighting conditions using their colors. To reduce the lighting effects, a reference image is first selected for building the mapping function between the current frame and the reference image. With this mapping function, the color distortions between frames can be reduced to minimum. In addition to lighting changes, the effect of sun light will make the vehicle window become white and lead to the errors of vehicle classification. To reduce this effect, a window-removing task is then applied for making vehicle pixels with the same color more concentrated on the foreground region. Then, vehicles can be more accurately classified to their categories even though strong sun light casts on them. To tackle the confusion problem that some vehicle colors are too similar, e.g., “deep-blue” and “deepgreen”, a novel tree-based classifier is then designed for classifying vehicles to more detailed labels. Experimental results have proved that the proposed method is a robust, accurate, and powerful tool for vehicle classification.

Boosted road sign detection and recognition

This paper presents a boosted system to detect and recognize roads signs from videos. The system first uses the Adaboost algorithm to learn the visual characteristics of road sign. Then, a cascaded structure is then used to detect road signs from videos in real time. After detection, a rectification process is then applied for rectifying different skewed road signs into a normal one. Then, its all embedded texts can be more accurately recognized using their distance maps. On the map, a weighting function is used to balance the importance between a road signpsilas inner and outer feature so that its embedded characters can be more accurately recognized. Experimental results have proved the superiority of the proposed method in road sign recognition.

Edge-based Lane Change Detection and its Application to Suspicious Driving Behavior Analysis

This paper presents a novel edge labeling scheme for detecting lanes from videos in real time. Firstly, pairs of edge pixels with different edge types are grouped using the labeling technique. Then, different lane hypotheses can be generated for lane modeling. Then, a lane geometrical constraint is derived from the pinhole camera geometry for filtering out impossible lane hypotheses. Since the constraint is invariant to shadows and lighting changes, each desired lane can be robustly detected even though different occlusions and shadows are included in the analyzed scenes. After filtering, a kernel-based modeling technique is then proposed for modeling different lane properties. With the modeling, different lanes can be effectively detected and tracked even though they are fragmented into pieces of segments or occluded by shadows. The proposed scheme works very well to analyze lane conditions with night vision. With the lane information, different dangerous driving behaviors like lane departure can be directly analyzed from road scenes. Experimental results show that the proposed scheme is powerful in lane detection. The average accuracy rate of vehicle detection is 95%.

Occluded human body segmentation and its application to behavior analysis

This paper addresses the problem of occluded human segmentation and then uses its results for human behavior recognition. To tackle this ill-posed problem, a novel clustering scheme is proposed for constructing a model space for posture classification. Then, a model-driven approach is proposed for separating an occluded region to individual objects. For reducing the model space, a particle filtering technique is then used for locating possible positions of each occluded object. Then, from the positions, the best model of each occluded object can be then selected using its distance maps. Then, a novel template re-projection technique is proposed for repairing an occluded object to a complete one. Due to occlusions, there will be many posture symbol converting errors in this representation. Instead of using a specific symbol, we code a posture using not only its best matched key posture but also its similarities among other key postures. With the matrix representation, different actions can be more robustly and effectively matched by comparing their KL distance.

Abnormal Event Analysis Using Patching Matching and Concentric Features

This paper proposes a novel patch-based approach for abnormal event detection from a mobile camera using concentric features. It is very different from traditional methods which require the cameras being static for well foreground object detection. Two stages are included in this system i.e., training and detection, for scene representation and exceptional change detection of important objects like paintings or antiques. Firstly, at the training stage, a novel scene representation scheme is proposed for large-scale surveillance using a set of corners and key frames. Then, at the detection stage, a novel patch matching scheme is proposed for efficient scene searching and comparison. The scheme reduces the time complexity of matching not only from search space but also feature dimension in similarity matching. Thus, desired scenes can be obtained extremely fast. After that, a spider-web structure is proposed for missing object detection even though there are large camera movements between any two adjacent frames. Experimental results prove that our proposed system is efficient, robust, and superior in missing object detection and abnormal event analysis.