Browsing by Author "Hsu, Chun-Chieh"

Now showing 1 - 11 of 11

2D Histogram-based player localization in broadcast volleyball videos
(2016-06)
Hsu, Chun-Chieh
;
Chen, Hua-Tsung
;
Chou, Chien-Li
;
Lee, Suh-Yin
;
資訊工程學系
Player location is one of the most informative cues for obtaining tactics arrangement and collecting descriptive game statistics. However, state-of-the-art supervised learning-based methods for player localization require a large amount of labeled training data. Hence, the development of automatic systems for player localization becomes indispensable. Volleyball games reach a huge audience base and contain a variety of tactical strategies, necessitating the implementation of systems for inferring tactics and analyzing formations automatically. Therefore, a novel 2D histogram-based player localization method capable of locating players with occlusions is developed and presented in this paper. The proposed system is able to automatically detect the court lines for camera calibration, extract players by calculating both x and y histograms of extracted player masks, and visualize the team formations on real-world court model. The experiments on broadcast volleyball videos demonstrate efficient and effective results against a traditional object segmentation method (connected component analysis) and a supervised learning approach utilizing histogram of oriented gradient features.
A NOVEL VIDEO ANNOTATION FRAMEWORK USING NEAR-DUPLICATE SEGMENT DETECTION
(2015)
Chou, Chien-Li
;
Chen, Hua-Tsung
;
Hsu, Chun-Chieh
;
Lee, Suh-Yin
;
資訊工程學系
The traditional video annotation approaches focus on annotating keyframes, shots, or the whole video with semantic keywords. However, the extractions of keyframes and shots lack of semantic meanings, and it is hard to use a few keywords to describe a video by using multiple topics. Therefore, we propose a novel video annotation framework using near-duplicate segment detection not only to preserve but also to purify the semantic meanings of target annotation units. A hierarchical near-duplicate segment detection method is proposed to efficiently localize near-duplicate segments in frame-level. Videos containing near-duplicate segments are clustered and keyword distributions of clusters are analyzed. Finally, the keywords ranked according to keyword distribution scores are annotated onto the obtained annotation units. Comprehensive experiments demonstrate the effectiveness of the proposed video annotation framework and near-duplicate segment detection method.
A System for Front Vehicle Detection and Fast Indexing Using a Single Car-Mounted Camera
(2015-01-01)
Hsu, Chun-Chieh
;
Wu, Yi-Chien
;
Chen, Hua-Tsung
;
Lee, Suh-Yin
;
交大名義發表
Video safety becomes a recently arising issue in the automobile industry due to the rapid expansion of car ownership worldwide. Increasing driver assistance systems have been developed for security and accident prevention by continuously monitoring the vehicle surroundings and the driving behaviors to detect potential problems in advance. Since most traffic/criminal accidents are caused by vehicles, recognizing and tracking the fleeing vehicles on the roads are of vital importance. In this paper, we propose a vision-based front vehicle detection and fast indexing system using a single car-mounted camera, which tends to be widespread deployed. First, front vehicle candidates are detected by a cascade of Gentle Adaboost classifiers utilizing Histogram of Oriented Gradients features. To improve the detection accuracy, symmetry of taillights are further employed to filter out the false detected candidates. Once a vehicle is detected, the GPS locations are reported to a central server, so that a database of detections can be maintained and can be provided to the related departments for timely retrieving and quickly indexing the searched vehicles on the maps. Experiments conducted on real videos demonstrate that satisfactory results can be obtained by the proposed detection scheme.
Computer-assisted Billiard Self-Training Using Intelligent Glasses
(2017-01-01)
Hsu, Chun-Chieh
;
Tsai, Hou-Chun
;
Chen, Hua-Tsung
;
Tsai, Wen-Jiin
;
Lee, Suh-Yin
Self-training plays an important role in sports exercise. However, if not under the instruction of a coach, it would be ineffective for most amateurs or inexperienced players to exercise on their own. Therefore, establishing computer-assisted training systems for sports exercise is a recently emerging topic. In this paper, we propose a billiard self-training system, which aims at improving billiard players' performance by utilizing intelligent glasses as a wearable camera and displayer. The proposed system is able to automatically analyze user-captured images of the billiard table from multiple views and display the ball configurations on a virtual top-view table. Enriched visual presentation can be provided to give the practitioner a further sight into the game. The experiments conducted on sixteen sets of different ball configurations show promising results.
Computer-assisted yoga training system
(2018-09-01)
Chen, Hua-Tsung
;
He, Yu-Zhen
;
Hsu, Chun-Chieh
;
資訊工程學系
;
Department of Computer Science
Self-training is essential in sports exercise. However, without the instruction of a coach, a practitioner may progress to a limited extent. Improper postures may even cause serious harm to muscles and ligaments of the body. Hence, the development of computer-assisted self-training systems for sports exercise is a recently emerging research topic. In this paper, we propose a yoga self-training system, which aims at instructing the practitioner to perform yoga poses correctly, assisting in rectifying poor postures, and preventing injury. Integrating computer vision techniques, the proposed system analyzes the practitioner's posture from both front and side views by extracting the body contour, skeleton, dominant axes, and feature points. Then, based on the domain knowledge of yoga training, visualized instructions for posture rectification are presented so that the practitioner can easily understand how to adjust his/her posture. Experiments on twelve yoga poses performed by different practitioners validate the feasibility of the proposed system in yoga training.
Daytime Preceding Vehicle Brake Light Detection Using Monocular Vision
(2016-01-01)
Chen, Hua-Tsung
;
Wu, Yi-Chien
;
Hsu, Chun-Chieh
;
資訊工程學系
;
Department of Computer Science
Advanced vehicle safety is a recently emerging issue appealed from the explosive population of car owners. Increasing driver assistance systems have been developed for warning drivers of potential hazards by analyzing the surroundings with sensors and/or cameras. Issuing vehicle deceleration and potential collision, brake lights are particularly important warning signals, allowing of no neglect. In this paper, we propose a vision-based daytime brake light detection system using a driving video recorder, which tends to be widespread used. At daytime, visual features, motions, and appearances of vehicles are highly visible. However, brake lights, on the contrary, are hard to notice due to low contrast between the brakes lights and environments. Without the significant characteristic of light scattering as at night, the proposed system extracts preceding vehicles with taillight symmetry verification, and then integrates both luminance and radial symmetry features to detect brake lights. A detection refinement process using temporal information is also employed for miss recovery. Experiments are conducted on a test data set collected by front-mounted driving video recorders, and the results verify that the proposed system can effectively detect brake lights at daytime, showing its good feasibility in real-world environments.
Fast Multi-View People Localization Using a Torso-High Reference Plane
(2017-01-01)
Hsu, Chun-Chieh
;
Chen, Hua-Tsung
;
Tsai, Wen-Jiin
;
Lee, Suh-Yin
;
交大名義發表
People locations bring rich information for a wide spectrum of applications in intelligent video surveillance systems. In addition to localization accuracy, computational efficiency is another significant issue to be highly concerned in people localization. As an essential early stage, people localization has to be accomplished in a very short time, enabling further semantic analysis. However, most state-of-the-art people localization methods pay little attention to computational efficiency. Hence, in this paper we propose an effective and efficient multi-view people localization scheme with several acceleration mechanisms. First, a torso-high reference plane is introduced since in general the torso part (after foreground segmentation) is more intact and stable than the other parts of a human body, and thus can predict potential people locations more reliably. Then, a novel and computationally efficient bitwise-operation scheme is proposed to predict people locations at the intersection regions of foreground line samples from multiple views. After rule-based validation, people locations can be accurately obtained and visualized on a real world plane. Experiments on multi-view surveillance videos not only validate the high accuracy of the proposed method in locating people under crowded scenes with serious occlusions, but also demonstrate an outstanding computational speed.
NEAR-DUPLICATE VIDEO RETRIEVAL BY USING PATTERN-BASED PREFIX TREE AND TEMPORAL RELATION FOREST
(2014-01-01)
Chou, Chien-Li
;
Chen, Hua-Tsung
;
Hsu, Chun-Chieh
;
Ho, Chien-Peng
;
Lee, Suh-Yin
With the explosive growth of the social multimedia sharing, copyright protection and search result refinement are always the critical issues for the service operators. To resolve the problems, content-based near-duplicate video retrieval is developed in recent years. In this paper, we construct a condensed Pattern-based Prefix tree (PP-tree) to index the patterns of reference videos for fast retrieval. To calculate how likely a query video and a reference video are near-duplicates, a novel algorithm for discovering the temporal relations among patterns is proposed. Comprehensive experiments on public datasets are conducted to verify the effectiveness and efficiency of the proposed method. Experimental results show that the proposed near-duplicate video retrieval approach outperforms the state-of-the-art approaches in terms of precision, recall, and execution time.
SERVE RECEIVE-TO-ATTACK PERIOD EXTRACTION AND HISTOGRAM-BASED PLAYER LOCALIZATION IN BROADCAST VOLLEYBALL VIDEOS
(2013)
Hsu, Chun-Chieh
;
Chen, Hua-Tsung
;
Chou, Chien-Li
;
Lee, Suh-Yin
;
Ho, Chien-Peng
Serve receive-to-attack (SR2A) is the most principal way to gain points in volleyball games. In addition, the positions of players on the court reveal informative clues about both offensive and defensive formations. Hence, in this paper, we propose an effective system capable of extracting SR2A periods from long broadcast volleyball videos (taken from a pan-tilt-zoom camera), and then locating the players using a novel histogram-based approach. Enriched visual presentation can be provided to give the audience or professional players/coaches a further sight into the game. The proposed system consists of four major processing modules, including court detection, SR2A period extraction, camera calibration, and histogram-based player localization. The experiments conducted on broadcast volleyball videos show promising results.
Trajectory Based Jump Pattern Recognition in Broadcast Volleyball Videos
(2014-01-01)
Hsu, Chun-Chieh
;
Chen, Hua-Tsung
;
Chou, Chien-Li
;
Ho, Chien-Peng
;
Lee, Suh-Yin
Jump actions are typically accompanied by spiking and imply significant events in volleyball matches. In this paper, we propose an effective system capable of jump pattern recognition in player moving trajectories from long broadcast volleyball videos. First, the entire video is segmented into clips of rallies by shot segmentation and whistle detection. Then, camera calibration is adopted to find the correspondence between coordinates in the video frames and real-world coordinates. With the homographic transformation matrix computed, real-world player moving trajectories can be derived by a sequence of tracked player locations in video frames. Jump patterns are recognized from the player moving trajectory by using a sliding window scheme with physics-based validation and context constraint. Finally, the jump locations can be estimated and jump tracks can be separated from the planar moving tracks. The experiments conducted on broadcast volleyball videos show promising results.
Vision-Based Road Bump Detection Using a Front-Mounted Car Camcorder
(2014-01-01)
Chen, Hua-Tsung
;
Lai, Chun-Yu
;
Hsu, Chun-Chieh
;
Lee, Suh-Yin
;
Lin, Bao-Shuh Paul
Advanced vehicle safety is a recently emerging issue, appealed from the rapidly explosive population of car owners. Increasing driver assistance systems have been designed for warning drivers of what should be noticed by analyzing surrounding environments with sensors and/or cameras. As one of the hazard road conditions, road bumps not only damage vehicles but also cause serious danger, especially at night or under poor lighting conditions. In this paper we propose a vision-based road bump detection system using a front-mounted car camcorder, which tends to be widespread deployed. First, the input video is transformed into a time-sliced image, which is a condensed video representation. Consequently, we estimate the vertical motion of the vehicle based on the time-sliced image and infer the existence of road bumps. Once a bump is detected, the location fix obtained from GPS is reported to a central server, so that the other vehicles can receive warnings when approaching the detected bumpy regions. Encouraging experimental results show that the proposed system can detect road bumps efficiently and effectively. It can be expected that traffic security will be significantly promoted through the mutually beneficial mechanism that a driver who is willing to report the bumps he/she meets can receive warnings issued from others as well.