Earticle

This paper presents a hierarchical deformable model for robust human face detection, especially with occlusions and under low resolution. By parsing, we mean inferring the parse tree (a configuration of the proposed hierarchical model) for each face instance. In modeling, a three-layer hierarchical model is built consisting of six nodes. For each node, an active basis model is trained, and their spatial relations such as relative locations and scales are modeled using Gaussian distributions. In computing, we run the learned active basis models on testing images to obtain bottom-up hypotheses, followed by explicitly testing the compatible relations among those hypotheses to do verification and construct the parse tree in a top-down manner. In experiments, we test our approach on CMU+MIT face test set with improved performance obtained.

Bacterial Δ$\^$5/-3-ketosteroid isomerase (KSI) is one of the most proficient enzymes catalyzing the isomerization of a variety of Δ$\^$5/-ketosteroids to Δ$^4$-ketosteroids at a diffusion-controlled rate. Because of the simplicity of the reaction, the enzyme mechanism has been intensively studied as a prototype to understand enzyme-catalyzed C-H bond cleavage.(omitted)

본 논문에서 제안하는 알고리즘은 수중에서 협대역 지속파 능동 소나를 이용하여 표적 반향음을 탐지하는 문제를 다루고 있다. 능동 소나에서 표적 탐지를 위해 방사한 핑 신호는 주변의 많은 산란체에 의해 반사되는 신호를 만들어내며, 이를 잔향이라 한다. 본 논문에서 제안하는 알고리즘은 잔향 환경에서 낮은 도플러의 표적 반향음을 탐지하는 것을 목표로 한다. 제안하는 알고리즘은 빔공간 다채널 비음수 행렬 분해 기법을 기반으로 하여 방위, 주파수, 시간 기저를 추정하며, 특히 기저를 두 개의 기저집단 -잔향음 기저집단과 반향음 기저집단으로 나누어 독립적으로 추정한다. 제안하는 알고리즘의 성능을 평가하기 위하여 합성된 잔향 신호를 이용하여 시뮬레이션을 진행하였으며, 시뮬레이션 결과 제안하는 알고리즘이 기존의 알고리즘에 비해 향상된 성능을 보이는 것을 확인할 수 있었다.

The proposed algorithm deals with a detection problem of target echo for narrow-band continuous-wave active sonar in the underwater environment in this paper. In the active sonar systems, ping signal emitted for target detection produces a signal that consists of multiple reflections by many scatterers around, which is called reverberation. The proposed algorithm aims to detect the low-Doppler target echo in the reverberant environment. The proposed algorithm estimates the bearing, frequency, and temporal bases based on beamspace-domain multichannel nonnegative matrix factorization. In particular, the bases are divided into two basis groups - the reverberation group and the echo group, then the basis groups are estimated independently. In order to evaluate the proposed algorithm, a simulation with synthesized reverberation was performed. The results show that the proposed algorithm has enhanced performance than the conventional algorithms.