• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.21 no.4
• /
• pp.11-18
• /
• 2007

A moving object detection algorithm for surveillance video is here proposed which employs background initialization based on two-stage median filtering and a background updating method based on state transition diagram. In the background initialization, the spatiotemporal similarity is measured in the subinterval. From the accumulated difference between the base frame and the other frames in a subinterval, the regions affected by moving objects are located. The median is applied over the subsequence in the subinterval in which regions share similarity. The outputs from each subinterval are filtered by a two-stage median filter. The background of every frame is updated by the suggested state transition diagram The object is detected by the difference between the current frame and the updated background. The proposed method showed good results even for busy, crowded sequences which included moving objects from the first frame.

• Journal of Broadcast Engineering
• /
• v.12 no.3
• /
• pp.289-292
• /
• 2007

A background initialization algorithm based on the spatiotemporal similarity measure in a motion tracking system is proposed. From the accumulated difference between the base frame and the other frames in a subinterval, the regions affected by moving objects are located. The median is applied over the subsequence in the subinterval in which co-located regions share the similarity. The outputs from each subinterval are filtered by second stage median filter. The proposed method showed good results even in the busy and crowded sequences where the real background does not exit.

• Proceedings of the IEEK Conference
• /
• 2004.08c
• /
• pp.810-813
• /
• 2004

In this paper, a background initialization for video surveillance proposed. The proposed algorithm is that the background images are sampled n frames during ${\Delta}t$ All Sampling frames are divided by $M{\times}N$ size block every frame. Average values of pixels for same location block of the sampling frames during ${\Delta}t$t are taken. then the maximum intensity $\alpha$ and the minimun intensity $\beta$ is obtained, respecticely. The intial by $M{\times}N$ size block, then average intensity $\eta$ of pixels for the block is obtained. If the average intensity $\eta$ is out of the initial range of the background image, it is decided the moving object image, and if the average intensity $\eta$ is included in the initial range of the background image. it is decided the background image. To examine the propriety of the proposed algorithm in this paper, the accuracy and robustness evaluation results for human and car in the indoor and outdoor enviroment. the error rate of the proposed method is less than the existing methods and the extraction rate of the proposed method is better than the existing methods.

• Proceedings of the Korean Institute of Communication Sciences Conference
• /
• 2004.07a
• /
• pp.223-223
• /
• 2004

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.1C
• /
• pp.44-53
• /
• 2005

In this paper, a improved moving object detection algorithm for stable performance of surveillance system in case of iterative moving in limited area and rapidly illuminance change in background scene is proposed. The proposed algorithm is that background scenes are sampled for initializing background image then the sampled fames are divided by block and sum of graylevel value for each block pixel was calculated, respectively. The initialization of background image is that background frame is respectively reconstructed with selecting only the maximum graylevel value and the minimum graylevel value of blocks located at same position between adjacent frames, then reference images of background are set by the reconstructed background images. Moving object detecting is that the current image frame is divided by block then sum of graylevel value for each block pixel is calculated. If the calculated value is out of graylevel range of the initialized two reference images, it is decided with moving objects block, otherwise it is decided background. The evaluated results is that the error rate of the proposed method is less than the error rate of the existing methods from $0.01{\%}$ to $20.33{\%}$ and the detection rate of the proposed method is better than the existing methods from $0.17{\%}\;to\;22.83{\%}$.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.6
• /
• pp.670-677
• /
• 2019

As the presence of background noise in acoustic signal degrades the performance of speech or acoustic event recognition, it is still challenging to extract noise-robust acoustic features from noisy signal. In this paper, we propose a combined structure of Wasserstein Generative Adversarial Network (WGAN) and MultiTask AutoEncoder (MTAE) as deep learning architecture that integrates the strength of MTAE and WGAN respectively such that it estimates not only noise but also speech features from noisy acoustic source. The proposed MTAE-WGAN structure is used to estimate speech signal and the residual noise by employing a gradient penalty and a weight initialization method for Leaky Rectified Linear Unit (LReLU) and Parametric ReLU (PReLU). The proposed MTAE-WGAN structure with the adopted gradient penalty loss function enhances the speech features and subsequently achieve substantial Phoneme Error Rate (PER) improvements over the stand-alone Deep Denoising Autoencoder (DDAE), MTAE, Redundant Convolutional Encoder-Decoder (R-CED) and Recurrent MTAE (RMTAE) models for robust speech recognition.

• 한국HCI학회:학술대회논문집
• /
• 2007.02a
• /
• pp.630-635
• /
• 2007

In this paper, we propose a system for multiple people tracking using fragment based histogram matching. Appearance model is based on IHLS color histogram which can be calculated efficiently using integral histogram representation. Since histograms will loss all spatial information, we define a fragment based region representation which retain spatial information, robust against occlusion and scale issue by using disparity information. Multiple people labeling is maintained by creating online appearance representation for each people detected in scene and calculating fragment vote map. Initialization is performed automatically from background segmentation step.

• Journal of KIISE:Computing Practices and Letters
• /
• v.9 no.6
• /
• pp.712-717
• /
• 2003

A novel multi-snake is proposed for efficient extraction of multi-object contours from a natural scene. An NTGST(noise-tolerant generalized symmetry transform) is used as a context-free attention operator to detect and locate multiple objects from a complex background and then the snake points are automatically initialized nearby the contour of each detected object using symmetry map of the NTGST before multiple snakes are introduced. These procedures solve the knotty subjects of automatic snake initialization and simultaneous extraction of multi-object contours in conventional snake algorithms. Because the snake points are initialized nearby the actual contour of each object, as close as possible, contours with high convexity and/or concavity can be easily extracted. The experimental results show that the proposed method can efficiently extract multi-object contours from a noisy and complex background of natural scenes.

• Asian Pacific Journal of Cancer Prevention
• /
• v.16 no.1
• /
• pp.77-81
• /
• 2015

Background: The aim of the study was to determine whether the expression of baseline phosphorus (P) and magnesium (Mg) levels were prognostic in terms of stage and overall survival (OS) in newly diagnosed non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) patients. Materials and Methods: Retrospectively, 130 patients were selected at the time of diagnosis oflung cancer (100 with NSCLC and 30 with SCLC), before the initialization of any chemo-radiotherapy. The median age was 67 (range 29-92). IA, IB, IIA, IIB, IIIA, IIIB and IV stages were present in 3, 4, 19, 6, 25, 8, and 65 patients, respectively. After centrifugation, the levels of serum P and Mg were measured using the nephelometric method/ photometry and evaluated before any type of treatment. Results: Higher than normal levels of P were found in 127/130 patients, while only four patients had elevated Mg serum values. In terms of Spearman test, higher P serum values correlated with either stage (rho=- 0.334, p<0.001) or OS (rho=-0.212, p=0.016). Additionally, a significant negative correlation of Mg serum levels was found with stage of disease (rho=-0.135, P=0.042). On multivariate cox-regression survival analysis, only stage (p<0.01), performance status (p<0.01) and P serum (p=0.045) showed a significant prognostic value. Conclusions: Our study indicated that pre-treatment P serum levels in lung cancer patients are higher than the normal range. Moreover, P and Mg serum levels are predictive of stage of disease. Along with stage and performance status, the P serum levels had also a significant impact on survival. This information may be important for stratifying patients to specific treatment protocols or intensifying their therapies. However, larger series are now needed to confirm our results.