• MALSORI
• /
• no.59
• /
• pp.27-51
• /
• 2006

For speech recognition systems, rejection function as well as decoding function is necessary to improve the reliability. There have been many research efforts on out-of-vocabulary word rejection, however, little attention has been paid on non-target sentence rejection. Recently pronunciation approaches using speech recognition increase the need for non-target sentence rejection to provide more accurate and robust results. In this paper, we proposed filler model method and word/phoneme detection ratio method to implement non-target sentence rejection system. We made performance evaluation of filler model along to word-level, phoneme-level, and sentence-level filler models respectively. We also perform the similar experiment using word-level and phoneme-level word/phoneme detection ratio method. For the performance evaluation, the minimized average of FAR and FRR is used for comparing the effectiveness of each method along with the number of words of given sentences. From the experimental results, we got to know that word-level method outperforms the other methods, and word-level filler mode shows slightly better results than that of word detection ratio method.

• Journal of Korea Multimedia Society
• /
• v.24 no.5
• /
• pp.659-666
• /
• 2021

With the continuous development of deep learning, human behavior recognition algorithms have achieved good results. However, in a multi-person recognition environment, the complex behavior environment poses a great challenge to the efficiency of recognition. To this end, this paper proposes a multi-person pose estimation model. First of all, the human detectors in the top-down framework mostly use the two-stage target detection model, which runs slow down. The single-stage YOLOv3 target detection model is used to effectively improve the running speed and the generalization of the model. Depth separable convolution, which further improves the speed of target detection and improves the model's ability to extract target proposed regions; Secondly, based on the feature pyramid network combined with context semantic information in the pose estimation model, the OHEM algorithm is used to solve difficult key point detection problems, and the accuracy of multi-person pose estimation is improved; Finally, the Euclidean distance is used to calculate the spatial distance between key points, to determine the similarity of postures in the frame, and to eliminate redundant postures.

• Journal of KIISE
• /
• v.42 no.1
• /
• pp.138-143
• /
• 2015

In this paper we develop a novel human action recognition system based on communication between an ear-mounted Bluetooth camera and an action recognition server to aid scene recognition for the blind. First, if the blind capture an image of a specific location using the ear-mounted camera, the captured image is transmitted to the recognition server using a smartphone that is synchronized with the camera. The recognition server sequentially performs human detection, object detection and action recognition by analyzing human poses. The recognized action information is retransmitted to the smartphone and the user can hear the action information through the text-to-speech (TTS). Experimental results using the proposed system showed a 60.7% action recognition performance on the test data captured in indoor and outdoor environments.

• Journal of Korea Multimedia Society
• /
• v.14 no.8
• /
• pp.992-999
• /
• 2011

This paper proposed a human body posture recognition program based on haar-like feature and hand activity detection. Its distinguishing features are the combination of face detection and motion detection. Firstly, the program uses the haar-like feature face detection to receive the location of human face. The haar-like feature is provided with the advantages of speed. It means the less amount of calculation the haar-like feature can exclude a large number of interference, and it can discriminate human face more accurately, and achieve the face position. Then the program uses the frame subtraction to achieve the position of human body motion. This method is provided with good performance of the motion detection. Afterwards, the program recognises the human body motion by calculating the relationship of the face position with the position of human body motion contour. By the test, we know that the recognition rate of this algorithm is more than 92%. The results show that, this algorithm can achieve the result quickly, and guarantee the exactitude of the result.

• Journal of Ocean Engineering and Technology
• /
• v.14 no.1
• /
• pp.80-87
• /
• 2000

Despite many combined research efforts outstanding needs exist to develop robust safety-estimation methods for large complex structures. This paper presents a practical damage identification scheme which can be applied to truss structures using only limited modal responses. firstly a theory of pattern recognition (PR) is described. Secondly existing damage-detection algorithms are outlined and a newly-derived algorithms for truss structures. Finally the feasibility of the proposed scheme is evaluated using numerical examples of plane truss structures.

• Journal of Korea Society of Industrial Information Systems
• /
• v.20 no.2
• /
• pp.57-63
• /
• 2015

Compared to other emotion recognition technology, facial expression recognition technology has the merit of non-contact, non-enforceable and convenience. In order to apply to a psychological robot, vision technology must be able to quickly and accurately extract the face region in the previous step of facial expression recognition. In this paper, we remove the background from any image using the YCbCr skin color technology, and use Haar-like Feature technology for robust face detection. We got the result of improved processing speed and robust face detection by removing the background from the input image.

• The KIPS Transactions:PartB
• /
• v.13B no.3 s.106
• /
• pp.315-322
• /
• 2006

Ant Colony System(ACS) is easily applicable to the traveling salesman problem(TSP) and it has demonstrated good performance on TSP. Recently, ACS has been emerged as the useful tool for the pattern recognition, feature extraction, and edge detection. The edge detection is wifely utilized in the area of document analysis, character recognition, and face recognition. However, the conventional operator-based edge detection approaches require additional postprocessing steps for the application. In the present study, in order to overcome this shortcoming, we have proposed the new ACS-based edge detection algorithm. The experimental results indicate that this proposed algorithm has the excellent performance in terms of robustness and flexibility.

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.6
• /
• pp.57-64
• /
• 1999

In this paper, we investigated the robust endpoint detection algorithm in noisy environment. A new feature parameter based on a discrete wavelet transform is proposed for word boundary detection of isolated utterances. The sum of standard deviation of wavelet coefficients in the third coarse and weighted first detailed scale is defined as a new feature parameter for endpoint detection. We then developed a new and robust endpoint detection algorithm using the feature found in the wavelet domain. For the performance evaluation, we evaluated the detection accuracy and the average recognition error rate due to endpoint detection in an HMM-based recognition system across several signal-to-noise ratios and noise conditions.

• Proceedings of the Korean Society for Emotion and Sensibility Conference
• /
• 2009.05a
• /
• pp.219-222
• /
• 2009

Rapid object recognition is one of the main stream research themes focusing to reveal how human recognizes object and interacts with environment in natural world. This field of study is of consequence in that it is highly important in evolutionary perspective to quickly see the external objects and judge their characteristics to plan future reactions. In this study, we investigated how human detect natural scene objects and categorize them in a limited time frame. We applied Magnetoencepahlogram (MEG) while participants were performing detection (e.g. object vs. texture) or basic-level categorization (e.g. cars vs. dogs) tasks to track the dynamic interaction in human brain for rapid object recognition process. The results revealed that detection and categorization involves different temporal and functional connections that correlated for the successful recognition process as a whole. These results imply that dynamics in the brain are important for our interaction with environment. The implication from this study can be further extended to investigate the effect of subconscious emotional factors on the dynamics of brain interactions during the rapid recognition process.

• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.12
• /
• pp.1226-1232
• /
• 2010

A door can serve as a feature for place classification and localization for navigation of a mobile robot in indoor environments. This paper proposes a door detection method based on the recognition of various door handles using the general Hough transform (GHT) and support vector machine (SVM). The contour and color histogram of a door handle extracted from the database are used in GHT and SVM, respectively. The door recognition scheme consists of four steps. The first step determines the region of interest (ROI) images defined by the color information and the environment around the door handle for stable recognition. In the second step, the door handle is recognized using the GHT method from the ROI image and the image patches are extracted from the position of the recognized door handle. In the third step, the extracted patch is classified whether it is the image patch of a door handle or not using the SVM classifier. The door position is probabilistically determined by the recognized door handle. Experimental results show that the proposed method can recognize various door handles and detect doors in a robust manner.