Browsing by Author "Huang, Ching-Chun"

Now showing 1 - 19 of 19

A Bayesian hierarchical detection framework for parking space detection
(2008)
Huang, Ching-Chun
;
Wang, Sheng-Jyh
;
Chang, Yao-Jen
;
Chen, Tsuhan
;
電子工程學系及電子研究所
hi this paper, a 3-layer Bayesian hierarchical detection framework (BHDF) is proposed for robust parking space detection. In practice, the challenges of the parking space detection problem come from luminance variations, inter-occlusions among cars, and occlusions caused by environmental obstacles. Instead of determining the status of parking spaces one by one, the proposed BHDF framework models the inter-occluded patterns as semantic knowledge and couple local classifiers with adjacency constraints to determine the status of parking spaces in a row-by-row manner. By applying the BHDF to the parking space detection problem, the available parking spaces and the labeling of parked cars can be achieved in a robust and efficient manner. Furthermore, this BHDF framework is generic enough to be used for various kinds of detection and segmentation applications.
A Bayesian Hierarchical Framework for Multitarget Labeling and Correspondence With Ghost Suppression Over Multicamera Surveillance System
(2012-01-01)
Huang, Ching-Chun
;
Wang, Sheng-Jyh
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In this paper, the main purpose is to locate, label, and correspond multiple targets with the capability of ghost suppression over a multicamera surveillance system. In practice, the challenges come from the unknown target number, the interocclusion among targets, and the ghost effect caused by geometric ambiguity. Instead of directly corresponding objects among different camera views, the proposed framework adopts a fusion-inference strategy. In the fusion stage, we formulate a posterior distribution to indicate the likelihood of having some moving targets at certain ground locations. Based on this distribution, a systematic approach is proposed to construct a rough scene model of the moving targets. In the inference stage, the scene model is inputted into a proposed Bayesian hierarchical detection framework, where the target labeling, target correspondence, and ghost removal are regarded as a unified optimization problem subject to 3-D scene priors, target priors, and foreground detection results. Moreover, some target priors, such as target height, target width, and the labeling results are iteratively refined based on an expectation-maximization (EM) mechanism to further boost system performance. Experiments over real videos verify that the proposed system can systematically determine the target number, efficiently label moving targets, precisely locate their 3-D locations, and effectively tackle the ghost problem. Note to Practitioners-As cooperative multicamera surveillance networks become more and more popular, the demand of multicamera information fusion and user-friendly representations becomes crucial. Motivated by the demand, this paper demonstrates a new system to efficiently integrate, summarize, and infer video messages from multiple client cameras. The ultimate goal is to provide a global view of the surveillance zone so that the managers in the control room may monitor the scene in an easier way. The main functions of the proposed system include the fusion of detection results from many client cameras, the summarization of consistent messages, and the inference of the target movement in the 3-D scene. In the near future, we will also take into consideration more information, like temporal clues, photometric clues, and object-level clues, in order to perform advanced scene analyses like abnormal behavior detection.
A CASCADED HIERARCHICAL FRAMEWORK FOR MOVING OBJECT DETECTION AND TRACKING
(2010)
Huang, Ching-Chun
;
Wang, Sheng-Jyh
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In this paper we propose a cascaded hierarchical framework for object detection and tracking. We claim that, by integrating both detection and tracking into a unified framework, the detection and tracking of multiple moving objects in a complicated environment become more robust. Under the proposed architecture, detection and tracking cooperate with each other. Based on the result of moving object detection, a dynamic model is adaptively maintained for object tracking. On the other hand, the updated dynamic model is used for both temporal prior propagation of object labels and the update of foreground/background models, which step further to help the detection of moving objects. The experiments show accurate results can be obtained under situations with foreground/background appearance ambiguity, camera shaking, and object occlusion.
A Hierarchical Bayesian Generation Framework for Vacant Parking Space Detection
(2010-12-01)
Huang, Ching-Chun
;
Wang, Sheng-Jyh
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In this paper, from the viewpoint of scene understanding, a three-layer Bayesian hierarchical framework (BHF) is proposed for robust vacant parking space detection. In practice, the challenges of vacant parking space inference come from dramatic luminance variations, shadow effect, perspective distortion, and the inter-occlusion among vehicles. By using a hidden labeling layer between an observation layer and a scene layer, the BHF provides a systematic generative structure to model these variations. In the proposed BHF, the problem of luminance variations is treated as a color classification problem and is tackled via a classification process from the observation layer to the labeling layer, while the occlusion pattern, perspective distortion, and shadow effect are well modeled by the relationships between the scene layer and the labeling layer. With the BHF scheme, the detection of vacant parking spaces and the labeling of scene status are regarded as a unified Bayesian optimization problem subject to a shadow generation model, an occlusion generation model, and an object classification model. The system accuracy was evaluated by using outdoor parking lot videos captured from morning to evening. Experimental results showed that the proposed framework can systematically determine the vacant space number, efficiently label ground and car regions, precisely locate the shadowed regions, and effectively tackle the problem of luminance variations.
A Novel Massive Deployment Solution Based on the Peer-to-Peer Protocol
(2019-01-02)
Shiau, Steven J. H.
;
Huang, Yu-Chiang
;
Yen, Ching-Hsuan
;
Tsai, Yu-Chin
;
Sun, Chen-Kai
The BitTorrent (BT) is a peer-to-peer (P2P) file sharing protocol that was developed approximately 20 years ago, is becoming increasingly popular, and has been widely accepted. The BT-based mass deployment system can be used to improve performance and scalability that cannot be achieved by the unicasting, broadcasting, and multicasting protocols. However, when the BT-based system is applied in massive deployments, a major issue related to insufficient temporary storage space to store the whole system image before deploying needs to be resolved. Such problems arose because the system is deployed to the disk space, meaning that it cannot be used for temporary storage. Therefore, a novel BT-based solution that can remove the limitations caused by the insufficient temporary storage issue is proposed. The BT-based mass deployment system was designed by using the file system blocks transferring (FSBT) mechanism. The receiver of the FSBT mechanism can obtain the blocks of the file system from other peers. Then, those blocks will be written directly to raw disks or partitions. The sender of the FSBT mechanism can read the blocks of file systems directly from raw disks or partitions. Then, the blocks can be sent to other peers. This approach solves the insufficient temporary storage issue. The novel BT-based mass deployment system was tested and verified for the configuration consisting of at most 32 personal computers (PCs). To demonstrate the achievable performance of the novel BT-based system, comparisons were made between the novel program and the traditional multicast solutions, as well as other solutions for mass deployment. The proposed BT solution can be much faster than the multicast solution when deploying 11 machines or more. The experimental results demonstrated the feasibility and superior performance of the proposed system. Furthermore, performance comparisons of the proposed BT-based mass deployment system versus other solutions demonstrated the feasibility and efficiency of the proposed solution. In the future, the BT parameters can be further optimized, and the simultaneous read and write features can be implemented to improve the deployment performance. In addition, the BT-based mass deployment system can serve as the basis for the development of other mass deployment systems.
A Surface-based Vacant Space Detection for an Intelligent Parking Lot
(2012)
Huang, Ching-Chun
;
Dai, Yu-Shu
;
Wang, Sheng-Jyh
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
We proposed a surface-based vacant parking space detection system. Unlike many car-oriented or space-oriented methods, the proposed system is parking-lot-oriented. In the system, we treat the whole parking lot as a structure consisting of plentiful surfaces. A surface-based hierarchical framework is then proposed to integrate the 3-D scene information with the patch-based image observation for the inference of vacant space. To be robust, the feature vector of each image patch is extracted based on the Histogram of Oriented Gradients (HOG) approach. By incorporating these texture features into the proposed probabilistic models, we could systematically infer the optimal hypothesis of parking statuses while dealing with occlusion effect, shadow effect, perspective distortion, and fluctuation of lighting condition in both day time and night time.
Domain Adaptation Meets Disentangled Representation Learning and Style Transfer
(2019-01-01)
Vu-Hoang Tran
;
Huang, Ching-Chun
;
資訊工程學系
;
Department of Computer Science
In this paper, we face the challenges of unsupervised domain adaptation and propose a novel three-in-one framework where three tasks - domain adaptation, disentangled representation, and style transfer are considered simultaneously. Firstly, the learned features are disentangled into common parts and specific parts. The common parts represent the transferrable features, which are suitable for domain adaptation with less negative transfer. Conversely, the specific parts characterize the unique style of each individual domain. Based on this, the new concept of feature exchange across domains, which can not only enhance the transferability of common features but also be useful for image style transfer, is introduced. These designs allow us to introduce five types of training objectives to realize the three challenging tasks at the same time. The experimental results show that our architecture can be adaptive well to full transfer learning and partial transfer learning upon a well-learned disentangled representation. Besides, the trained network also demonstrates high potential to generate style-transferred images.
Domain Adaptation With Foreground/Background Cues and Gated Discriminators
(2020-07-01)
Lin, Yong-Xiang
;
Tan, Daniel Stanley
;
Chen, Yung-Yao
;
Huang, Ching-Chun
;
Hua, Kai-Lung
Self-driving cars leverage on semantic segmentation to understand an urban scene. However, it is costly to collect segmentation labels, thus, synthetic datasets are used to train segmentation models. Unfortunately, the synthetic to real domain shift causes these models to perform poorly. Prior works use adversarial training to align features of both synthetic and real-world images. We observe that background objects tend to be similar across domains, while foreground objects tend to have more variations. Using this insight, we propose an adaptation method that uses foreground and background cues and adapt them separately. We also propose a mask-aware gated discriminator that learns soft masks from the input foreground and background masks instead of naively performing binary masking that immediately removes information outside of the predicted masks. We evaluate our method on two different datasets and show that our method outperforms several state-of-the-art baselines, which verifies the effectiveness of our approach.
Dynamic Sampling Rate Adjustment for Compressive Spectrum Sensing over Cognitive Radio Network
(2012-04)
Huang, Ching-Chun
;
Wang, Li-Chun
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In this paper, a dynamic sampling rate adjustment scheme is proposed for compressive spectrum sensing in cognitive radio network. Nowadays, compressive sensing (CS) has been proposed with a revolutionary idea to sense the sparse spectrum by using a lower sampling rate. However, many methods for compressive spectrum sensing assume that the sparse level is static and a fixed compressive sampling rate is applied over time. To adapt to time-varying sparse levels and adjust the sampling rate, we proposed to model sparse levels as a dynamic system and treat the dynamic rate selection as a tracking problem. By introducing the Sequential Monte Carlo (SMC) algorithm into a distributed compressive spectrum sensing framework, we could not only track the optimal sampling rate but determine the unoccupied channels accurately in a unified method.
Helpful or unhelpful? Self-affirmation on challenge-confronting tendencies for students who fear being laughed at
(2016-01)
Liu, Cheng-Hong
;
Huang, Po-Sheng
;
Chang, Jen-Ho
;
Lin, Ching-Yi
;
Huang, Ching-Chun
Studies suggest that self-affirmation interventions may help increase students\' tendencies to confront challenges. However, we posited that self-affirmation cannot help students with higher levels of fear of being laughed at (gelotophobia) confront challenges because it may lead them to accept negative information (feelings of threat) about their ability and decrease their perceptions of self-ability. In support of this hypothesis, Experiment 1 found that being attribute-affirmed enhanced the tendencies to confront a challenging task for junior high school students with low gelotophobia, but not for those with high gelotophobia. Experiment 2 revealed that being value-affirmed increased challenge-confronting tendencies for senior high school students with low gelotophobia. However, it decreased high gelotophobia students\' challenge-confronting tendencies because it caused them to feel less capable of performing the task. Thus, self-affirmation, which may decrease high gelotophobia students\' self-perceived ability in challenging situations, may be unhelpful or even harmful in motivating them to confront challenges. (C) 2015 Elsevier Inc. All rights reserved.
High-speed data-plane packet aggregation and disaggregation by P4 switches
(2019-09-15)
Wang, Shie-Yuan
;
Wu, Chia-Ming
;
Lin, Yi-Bing
;
Huang, Ching-Chun
;
資訊工程學系
In this paper, we propose novel approaches that utilize the header manipulations of the P4 (Programming Protocol-Independent Packet Processor) switches to aggregate small IoT packets into a large one, transmit it over a network, and then disaggregate it back to the original small packets, all in the data plane of the hardware P4 switch to provide high throughputs. Packet aggregation and disaggregation provide many important benefits and have been proposed and performed in the past. However, most existing approaches perform packet aggregation and disaggregation in the control plane either by the switch CPU or by the server CPU, resulting in low throughputs. Our work is the first work that designs and implements packet aggregation and disaggregation purely in the pipelines of the switching ASIC. In this paper, we present the design and implementation of our approaches, their measured throughputs, and the insights that we have obtained from this pioneering work.
IUML: INCEPTION U-NET BASED MULTI-TASK LEARNING FOR DENSITY LEVEL CLASSIFICATION AND CROWD DENSITY ESTIMATION
(2019-01-01)
Huynh, Van-Su
;
Vu-Hoang Tran
;
Huang, Ching-Chun
;
交大名義發表
;
National Chiao Tung University
Nowadays, image-based people counting is an essential technique for public safety management. However, this work is still extremely challenging due to many kinds of scale issues caused by different congested scenes, different viewing points, different image sizes, and different density levels. In this paper, we proposed a CNNs-based framework for people counting and crowd density map estimation with the consideration of the scale problems. First, we introduced an encoder-decoder architecture, which is composed of Inception modules to learn the multi-scale feature representations. Besides, to be adaptive to image resolution, a multi-loss setting over different resolutions of density maps is designed for network training. Second, we apply multi-task learning to learn the joint features for the density map estimation task and the density level classification task. This helps to enhance the feature generality under different scenes. Finally, by adopting the U-net architecture, the encoder and decoder features are then fused to generate high-resolution density maps. The efficacy of the proposed method is evaluated in the extensive experiments by quantifying the counting performance through multiple evaluation criteria.
MOBILE LEARNING SYSTEM WITH CONTEXT-AWARE INTERACTIONS AND POINT-OF-INTEREST UNDERSTANDING
(2018-01-01)
Chen, Oscal Tzyh-Chiang
;
Yu, Sung-Nien
;
Huang, Ching-Chun
;
Lee, Huang-Chen
;
Hsueh, Yu-Ling
To explore museums and art galleries, visitors usually follow the routes and regulations that may limit learning interest and efficiency. How to improve and assess the learning effect needs to be well addressed. Accordingly, we have developed a mobile learning system which can effectively offer visitors with attention tracking, interest cultivation, and immersive engagement. This work adopts the commercially-available smart glass accompanied with an additional camera, a Photoplethysmography (PPG) sensor module, and a smart phone or tablet to fulfill the integrated computations from multiple sensors. Based on these devices, the proposed system includes (1) computation of visual focal points, (2) attention estimation via heart rate variability, (3) indoor positioning, (4) derivation of user preferences and behaviors, and (5) intimate interactions of mobile learning. The experimental results reveal that the proposed mobile learning system can effectively increase user's interest, personality, and learning effect via contextaware interactions and point-of-interest understanding.
MOVING TARGETS LABELING AND CORRESPONDENCE OVER MULTI-CAMERA SURVEILLANCE SYSTEM BASED ON MARKOV NETWORK
(2009)
Huang, Ching-Chun
;
Wang, Sheng-Jyh
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In this paper, we propose an efficient way to simultaneously label and map targets over a multi-camera surveillance system. In the system, we first fuse the detection results from multiple cameras into a posterior distribution. This distribution indicates the likelihood of having some moving targets on the ground plane. Based on the distribution, isolated targets, together with their 3-D positions, are identified in a sample-based manner, which combines Markov Chain Monte Carlo (MCMC), and Mean-Shift clustering. The induced 3-D scene information is further inputted into a 3-layer Bayesian hierarchical framework (BHF), which adopts a Markov network to deal with the object labeling and correspondence problems. In principle, labeling and correspondence are regarded as a unified optimal problem subject to 3-D scene prior, image color similarity, and detection results. The experiments show that accurate results can be gotten even under situations with severe occlusion.
Online Self-learning for Smart HVAC Control
(2019-01-01)
Tzu-Yin Chao
;
Manh-Hung Nguyen
;
Huang, Ching-Chun
;
Liang, Chien-Cheng
;
Chung, Chen-Wu
In this paper, we introduce an online-learning method to model the property of an office building. Unlike conventional control methods where the building property is modeled via a simulator or through offline learning, our building model is adaptively updated according to the dynamic response of a real environment. Upon the building model for environment prediction, the proposed action agent can control the heating, ventilation, and air conditioning (HVAC) system in a smarter way by scheduling the temperature reference point. To online learn the model and improve the agent, two practical and seldom discussed issues are addressed. The first challenge is data bias where the collected initial training dataset can only partially reveal the statistical mapping between the control input and the environment response. Hence, the trained model may lack generalization. To overcome the data bias issue, a data augmentation method is proposed to embed physical logic in order to train a proper initial model. Next, an online learning process is introduced to update the model generality during the system operation phase. The second practical issue is the constraints on agent exploration for discovering unknown data samples. During the business hours, to comfort employees, a control agent is not allowed to explore the possible controlling space randomly. To balance data collection and control stability, we introduce a hybrid control strategy that considers both the human control rule and the agent action. A confidence score of the agent model is also automatically estimated to determine a suitable control strategy finally. Our experiments have realized in an office building. The results outperform conventional methods and show its superior in terms of control stability.
Robust 3D Skeleton Tracking based on OpenPose and a Probabilistic Tracking Framework
(2019-01-01)
Huang, Ching-Chun
;
Nguyen, Manh Hung
;
資訊工程學系
;
Department of Computer Science
For human skeleton tracking, conventional methods may directly rely on the Kinect built-in function to extract skeleton joints and realize the tracking function. However, the left-right confusion problem and the self-occlusion problem would make the Kinect joints unstable and incorrect. In this paper, we aim to solve these crucial issues and realize a reliable 3D skeleton tracking system via multiple Kinect cameras. First, the proposed method corrected the Kinect skeleton by referring to the OpenPose-extracted joints. Since OpenPose extracts the joints by analyzing image content, we can differentiate the front side and the back side and thus can correct the left-right confusion. Second, to alleviate the self-occlusion problem, we back-project the 2D OpenPose joints from multiple cameras and form an alternative 3D skeleton. The OpenPose reconstructed joints are treated as robust 3D anchors for multiple skeleton fusion. Last but not least, we introduce the inter-joint constraints into our skeleton tracking framework so that we can trace all joints simultaneously, make sure the skeleton movement consistent, and well maintain the length between neighboring joints. We evaluate our method with challenging actions. A practical system is also built for demonstration. The experimental results show the system can track skeleton stably without error propagation and vibration. The average localization error is also smaller than conventional methods.
SDN Soft Computing Application for Detecting Heavy Hitters
(2019-10-01)
Lin, Yi-Bing
;
Huang, Ching-Chun
;
Tsai, Shi-Chun
;
資訊工程學系
;
Department of Computer Science
To avoid distributed denial-of-service (DDoS) attacks or real-time transmission control protocol (TCP) incast in the software-defined networking (SDN) environment, the HashPipe algorithm was developed following the space-saving approach. Unfortunately, HashPipe implemented in the behavioral model (bmv2) cannot be directly executed at a real programming protocol-independent packet processor (P4) switch due to P4 pipeline limitation. Based on the Banzai machine model, this paper shows how to smartly utilize the Banzai atoms to develop HashPipe as a soft computing application in a real P4 switch. Then we propose an enhanced HashPipe algorithm that significantly improves the accuracy of the original HashPipe. The proposed heavy hitter detection is executed at the line-rate of the Tofino P4 switch with the highest process rate in the world.
The SDN Approach for the Aggregation/Disaggregation of Sensor Data
(2018-07-01)
Lin, Yi-Bing
;
Wang, Shie-Yuan
;
Huang, Ching-Chun
;
Wu, Chia-Ming
;
資訊工程學系
In many Internet of Things (IoT) applications, large numbers of small sensor data are delivered in the network, which may cause heavy traffics. To reduce the number of messages delivered from the sensor devices to the IoT server, a promising approach is to aggregate several small IoT messages into a large packet before they are delivered through the network. When the packets arrive at the destination, they are disaggregated into the original IoT messages. In the existing solutions, packet aggregation/disaggregation is performed by software at the server, which results in long delays and low throughputs. To resolve the above issue, this paper utilizes the programmable Software Defined Networking (SDN) switch to program quick packet aggregation and disaggregation. Specifically, we consider the Programming Protocol-Independent Packet Processor (P4) technology. We design and develop novel P4 programs for aggregation and disaggregation in commercial P4 switches. Our study indicates that packet aggregation can be achieved in a P4 switch with its line rate (without extra packet processing cost). On the other hand, to disaggregate a packet that combines N IoT messages, the processing time is about the same as processing N individual IoT messages. Our implementation conducts IoT message aggregation at the highest bit rate (100 Gbps) that has not been found in the literature. We further propose to provide a small buffer in the P4 switch to significantly reduce the processing power for disaggregating a packet.
Vacant Parking Space Detection Based on Plane-based Bayesian Hierarchical Framework
(2013-09-01)
Huang, Ching-Chun
;
Tai, Yu-Shu
;
Wang, Sheng-Jyh
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In this paper, we propose a vacant parking space detection system that operates day and night. In the daytime, the major challenges of the system include dramatic lighting variations, shadow effect, inter-object occlusion, and perspective distortion. In the nighttime, the major challenges include insufficient illumination and complicated lighting conditions. To overcome these problems, we propose a plane-based method which adopts a structural 3-D parking lot model consisting of plentiful planar surfaces. The plane-based 3-D scene model plays a key part in handling inter-object occlusion and perspective distortion. On the other hand, to alleviate the interference of unpredictable lighting changes and shadows, we propose a plane-based classification process. Moreover, by introducing a Bayesian hierarchical framework to integrate the 3-D model with the plane-based classification process, we systematically infer the parking status. Last, to overcome the insufficient illumination in the nighttime, we also introduce a preprocessing step to enhance image quality. The experimental results show that the proposed framework can achieve robust detection of vacant parking spaces in both daytime and nighttime.