Browsing by Author "Chou, Kuang-Pen"

Now showing 1 - 8 of 8

A METHOD TO ENHANCE THE DEEP LEARNING IN AN AERIAL IMAGE
(2017-01-01)
Chou, Kuang-Pen
;
Li, Dong-Lin
;
Lin, Wen-Chieh
;
Prasad, Mukesh
;
Lin, Chin-Teng
In this paper, we propose a kind of pre-processing method which can be applied to the depth learning method for the characteristics of aerial image. This method combines the color and spatial information to do the quick background filtering. In addition to increase execution speed, but also to reduce the rate of false positives
A Motor Imagery Based Brain-Computer Interface System via Swarm-Optimized Fuzzy Integral and Its Application
(2016)
Wu, Shang-Lin
;
Liu, Yu-Ting
;
Chou, Kuang-Pen
;
Lin, Yang-Yin
;
Lu, Jie
A brain-computer interface (BCI) system provides a convenient means of communication between the human brain and a computer, which is applied not only to healthy people but also for people that suffer from motor neuron diseases (MNDs). Motor imagery (MI) is one well-known basis for designing Electroencephalography (EEG)-based real-life BCI systems. However, EEG signals are often contaminated with severe noise and various uncertainties, imprecise and incomplete information streams. Therefore, this study proposes spectrum ensemble based on swam-optimized fuzzy integral for integrating decisions from sub-band classifiers that are established by a sub-band common spatial pattern (SBCSP) method. Firstly, the SBCSP effectively extracts features from EEG signals, and thereby the multiple linear discriminant analysis (MLDA) is employed during a MI classification task. Subsequently, particle swarm optimization (PSO) is used to regulate the subject-specific parameters for assigning optimal confidence levels for classifiers used in the fuzzy integral during the fuzzy fusion stage of the proposed system. Moreover, BCI systems usually tend to have complex architectures, be bulky in size, and require time-consuming processing. To overcome this drawback, a wireless and wearable EEG measurement system is investigated in this study. Finally, in our experimental result, the proposed system is found to produce significant improvement in terms of the receiver operating characteristic (ROC) curve. Furthermore, we demonstrate that a robotic arm can be reliably controlled using the proposed BCI system. This paper presents novel insights regarding the possibility of using the proposed MI-based BCI system in real-life applications.
A self-adaptive artificial bee colony algorithm with local search for TSK-type neuro-fuzzy system training
(2019-01-01)
Chou, Kuang-Pen
;
Lin, Chin-Teng
;
Lin, Wen-Chieh
;
資訊工程學系
;
Department of Computer Science
In this paper, we introduce a self-adaptive artificial bee colony (ABC) algorithm for learning the parameters of a Takagi-Sugeno-Kang-type (TSK-type) neuro-fuzzy system (NFS). The proposed NFS learns fuzzy rules for the premise part of the fuzzy system using an adaptive clustering method according to the input-output data at hand for establishing the network structure. All the free parameters in the NFS, including the premise and the following TSK-type consequent parameters, are optimized by the modified ABC (MABC) algorithm. Experiments involve two parts, including numerical optimization problems and dynamic system identification problems. In the first part of investigations, the proposed MABC compares to the standard ABC on mathematical optimization problems. In the remaining experiments, the performance of the proposed method is verified with other metaheuristic methods, including differential evolution (DE), genetic algorithm (GA), particle swarm optimization (PSO) and standard ABC, to evaluate the effectiveness and feasibility of the system. The simulation results show that the proposed method provides better approximation results than those obtained by competitors methods.
Block-based Feature Extraction Model for Early Fire Detection
(2017-01-01)
Chou, Kuang-Pen
;
Prasad, Mukesh
;
Gupta, Deepak
;
Sankar, Sharmi
;
Xu, Ting-Wei
Every year the fire disaster always causes a lot of casualties and property damage. Many researchers are involved in the study of related disaster prevention. Early warning systems and stable fire can significantly reduce the damage caused by fire. Many existing image-based early warning systems can perform well in a particular field. In this paper, we propose a general framework that can be applied in most realistic environments. The proposed system is based on a block-based feature extraction method, which analyses local information in separate regions leading to a reduction in computing data. Local features of fire block are extracted from the detailed characteristics of fire objects, which include fire color, fire source immobility, and disorder. Each local feature has high detection rate and filter out different false-positive cases. Global analysis with fire texture and non-moving properties are applied to further reduce false alarm rate. The proposed system is composed of algorithms with low computation. Through a series of experiments, it can be observed that Experimental results show that the proposed system has higher detection rate and low false alarm rate under various environment.
Deep Sparse Representation Classifier for facial recognition and detection system
(2019-07-01)
Cheng, Eric-Juwei
;
Chou, Kuang-Pen
;
Rajora, Shantanu
;
Jin, Bo-Hao
;
Tanveer, M.
This paper proposes a two-layer Convolutional Neural Network (CNN) to learn the high-level features which utilizes to the face identification via sparse representation. Feature extraction plays a vital role in real-world pattern recognition and classification tasks. The details description of the given input face image, significantly improve the performance of the facial recognition system. Sparse Representation Classifier (SRC) is a popular face classifier that sparsely represents the face image by a subset of training data, which is known as insensitive to the choice of feature space. The proposed method shows the performance improvement of SRC via a precisely selected feature exactor. The experimental results show that the proposed method outperforms other methods on given datasets. (C) 2019 Elsevier B.V. All rights reserved.
Driving Fatigue Prediction with Pre-Event Electroencephalography (EEG) via a Recurrent Fuzzy Neural Network
(2016)
Liu, Yu-Ting
;
Wu, Shang-Lin
;
Chou, Kuang-Pen
;
Lin, Yang-Yin
;
Lu, Jie
We propose an electroencephalography (EEG) prediction system based on a recurrent fuzzy neural network (RFNN) architecture to assess drivers\' fatigue degrees during a virtual-reality (VR) dynamic driving environment. Prediction of fatigue degrees is a crucial and arduous biomedical issue for driving safety, which has attracted growing attention of the research community in the recent past. Meanwhile, combined with the benefits of measuring EEG signals facilitates, many EEG-based brain-computer interfaces (BCIs) have been developed for use in real-time mental assessment. In the literature, EEG signals are severely blended with stochastic noise; therefore, the performance of BCIs is constrained by low resolution in recognition tasks. For this rationale, independent component analysis (ICA) is usually used to find a source mapping from original data that has been blended with unrelated artificial noise. However, the mechanism of ICA cannot be used in real-time BCI design. To overcome this bottleneck, the proposed system in this paper utilizes a recurrent self-evolving fuzzy neural work (RSEFNN) to increase memory capability for adaptive noise cancellation when assessing drivers\' mental states during a car driving task. The experimental results without the use of ICA procedure indicate that the proposed RSEFNN model remains superior performance compared with the state-of-thearts models.
Fast Deformable Model for Pedestrian Detection with Haar-like Features
(2017-01-01)
Chou, Kuang-Pen
;
Prasad, Mukesh
;
Puthal, Deepak
;
Chen, Ping-Hung
;
Vishwakarma, Dinesh Kumar
This paper proposes a novel Fast Deformable Model for Pedestrian Detection (FDMPD) to detect the pedestrians efficiently and accurately in the crowded environment. Despite of multiple detection methods available, detection becomes difficult due to variety of human postures and perspectives. The proposed study is divided into two parts. First part trains six Adaboost classifiers with Haar-like feature for different body parts (e.g., head, shoulders, and knees) to build the response feature maps. Second part uses these six response feature maps with full-body model to produce spatial deep features. The combined deep features are used as an input to SVM to judge the existence of pedestrian. As per the experiments conducted on the INRIA person dataset, the proposed FDMPD approach shows greater than 44.75 % improvement compared to other state-of-the-art methods in terms of efficiency and robustness.
Robust Feature-Based Automated Multi-View Human Action Recognition System
(2018-01-01)
Chou, Kuang-Pen
;
Prasad, Mukesh
;
Wu, Di
;
Sharma, Nabin
;
Li, Dong-Lin
Automated human action recognition has the potential to play an important role in public security, for example, in relation to the multiview surveillance videos taken in public places, such as train stations or airports. This paper compares three practical, reliable, and generic systems for multiview video-based human action recognition, namely, the nearest neighbor classifier, Gaussian mixture model classifier, and the nearest mean classifier. To describe the different actions performed in different views, view-invariant features are proposed to address multiview action recognition. These features are obtained by extracting the holistic features from different temporal scales which are modeled as points of interest which represent the global spatial-temporal distribution. Experiments and cross-data testing are conducted on the KTH, WEIZMANN, and MuHAVi datasets. The system does not need to be retrained when scenarios are changed which means the trained database can be applied in a wide variety of environments, such as view angle or background changes. The experiment results show that the proposed approach outperforms the existing methods on the KTH and WEIZMANN datasets.