백장운 교수

In this paper, we proposes a novel drowsiness detection algorithm using a camera near the dashboard. The proposed algorithm detects the driver's face in the image and estimates the landmarks in the face region. In order to detect the face, the proposed algorithm uses an AdaBoost classifier based on the Modified Census Transform features. And the proposed algorithm uses regressing Local Binary Features for face landmark detection. Eye states (closed, open) is determined by the value of Eye Aspect

First Name is required invalid characters Last Name is required invalid characters Email Address is required Invalid Email Address Invalid Email Address

First Name is required invalid characters Last Name is required invalid characters Email Address is required Invalid Email Address Invalid Email Address

Abstract As the risk of accidental explosions at ammunition storage or hydrogen charging station increases in populated area, it is needed to design the facilities against blast loading, particularly subjected to partially confined explosion. However, the partially confined explosion lacks experimental test data to efficiently design the facilities subjected to the potential threat, when compared to unconfined or confined explosion cases. As a fundamental study on partially confined explosion, t

This paper proposes a vision-based side vehicle detection for blind spot detection systems. The proposed algorithm uses a HoG cascade classifier in order to detect vehicles, and tracks the detected vehicles with Kalman filter. The proposed algorithm performs a periodical vehicle detection instead of every frame vehicle detection. And the proposed algorithm reduces the detecting image size by downscaling the original image and setting the region of interest where vehicles can exist. As a result,

First Name is required invalid characters Last Name is required invalid characters Email Address is required Invalid Email Address Invalid Email Address

In this paper, we proposes a novel tracking algorithm combining Kalman Filter with mean-shift. Kalman Filter predicts the vehicle position in the next frame. Mean-shift finds the best candidate which has maximum similarity with the tracked vehicle in the predicted area. Kalman Filter updates its state value of vehicle position with the position of the best candidate from the mean-shift tracker. As a result, the proposed algorithm tracks the vehicle without local maximum problem of mean-shift tra

Deep learning technology has been used to develop improved license plate recognition (LPR) systems. In particular, deep neural networks have brought significant improvements in the LPR system. However, deep neural networks are vulnerable to adversarial examples. In the existing LPR system, adversarial examples study specific spots that are easily identifiable by humans or require human feedback. In this paper, we propose a method of generating adversarial examples in the license plate, which has

This study examines various algorithms for computing the convex hull of a set of n points in a d-dimensional space. Convex hulls are fundamental in computational geometry and are applied in computer graphics, pattern recognition, and computational biology. Such convex hulls can also be useful in symmetry problems. For instance, when points are arranged symmetrically, the convex hull is also likely to be symmetrically shaped, which can be useful for object recognition in computer vision or patter

Abstract Recently, as a new precast concrete (PC) construction method for increasing economy and constructability, the PC double-beam system has been developed for factories or logistic centers, where construction duration is particularly important. In this study, half-scaled PC double beam–column connection was tested under gravity loading and cyclic lateral loading. The major test parameters included the use of the spliced PC column and the addition of reinforcement at the beam–column joint. I

This study presents a novel data-driven approach for generating spectrum-matched earthquake ground motions using physics-informed neural networks (PINNs). The methodology leverages real recorded earthquake data and employs singular value decomposition for dimensionality reduction, enabling the extraction of eigen motions that capture correlated temporal patterns. By combining PINNs with these eigen motions, spectrum matching is achieved with clear physical interpretability. The generated motions