• 로봇학회논문지
• /
• 제10권1호
• /
• pp.16-23
• /
• 2015

This paper presents a method of improving the pose recognition accuracy of objects by using Kinect sensor. First, by using the SURF algorithm, which is one of the most widely used local features point algorithms, we modify inner parameters of the algorithm for efficient object recognition. The proposed method is adjusting the distance between the box filter, modifying Hessian matrix, and eliminating improper key points. In the second, the object orientation is estimated based on the homography. Finally the novel approach of Auto-scaling method is proposed to improve accuracy of object pose estimation. The proposed algorithm is experimentally tested with objects in the plane and its effectiveness is validated.

• 한국전자통신학회논문지
• /
• 제17권1호
• /
• pp.111-118
• /
• 2022

컴퓨터 비전 연구에서 2차원 인체 자세는 매우 광범위한 연구 방향으로 특히 자세 추적과 행동 인식에서 유의미한 분야다. 인체 자세 표적 획득은 이미지에서 인체 목표를 정확히 찾는 방법을 연구하는 것이 핵심이며 인체 자세 인식은 인공지능(AI)에 적용하는 한편 일상생활에 활용되고 있어서 매우 중요한 연구의의가 있다. 인체 자세 인식 효과의 우수성의 기준은 인식 과정의 성공률과 정확도에 의해 결정된다. 본 연구의 인체 자세 인식에서는 딥러닝 전용 데이터셋인 MS COCO를 기반하여 인체를 17개의 키 포인트로 구분하였다. 다음으로 주요 특징에 대한 세분화 마스크(segmentation mask) 방법을 사용하여 인식률을 개선하였다. 최종적으로 신경망 모델을 설계하고 간단한 단계별 학습부터 효율적인 학습에 이르기까지 많은 수의 표본을 학습시키는 알고리즘을 제안하여 정확도를 향상할 수 있었다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2009년도 추계학술대회 논문집
• /
• pp.175-176
• /
• 2009

• 한국정보통신학회논문지
• /
• 제14권2호
• /
• pp.469-478
• /
• 2010

본 논문은 3차원 발 자세를 추정하기 위한 효과적 형상 기술자를 제안하였다. 처리 시간을 단축시키기 위하여 특수 제작된 3차원 발 모형을 2차원 투영하여 발 형상 데이터베이스를 구축하고, 3차원 자세 요약정보를 메타 정보로 추가한 2.5차원 영상 데이터베이스를 구성하였다. 그리고 특징 공간 크기가 작고 다른 형상 기술자에 비하여 자세 추정 성능이 뛰어난 수정된 Centroid Contour Distance를 제안하였다. 제안된 기술자의 성능을 분석하기 위하여, 검색 정확도와 시공간 복잡도를 계산하고 기존의 방식들과 비교하였다. 실험 결과를 통하여 제안된 기술자는 특징 추출 시간과 자세 추정 정확도면에서 기존의 방식들보다 효과적인 것으로 나타났다.

• Smart Structures and Systems
• /
• 제18권4호
• /
• pp.801-818
• /
• 2016

To estimate structural displacement, a visually servoed paired structured light system (ViSP) was proposed in previous studies. The ViSP is composed of two sides facing each other, each with one or two laser pointers, a 2-DOF manipulator, a camera, and a screen. By calculating the positions of the laser beams projected onto the screens and rotation angles of the manipulators, relative 6-DOF displacement between two sides can be estimated. Although the performance of the system has been verified through various simulations and experimental tests, it has a limitation that the accuracy of the displacement measurement depends on the alignment of the laser pointers. In deriving the kinematic equation of the ViSP, the laser pointers were assumed to be installed perfectly normal to the same side screen. In reality, however, this is very difficult to achieve due to installation errors. In other words, the pose of laser pointers should be calibrated carefully before measuring the displacement. To calibrate the initial pose of the laser pointers, a specially designed jig device is made and employed. Experimental tests have been performed to validate the performance of the proposed calibration method and the results show that the estimated displacement with the initial pose calibration increases the accuracy of the 6-DOF displacement estimation.

• 한국멀티미디어학회논문지
• /
• 제25권9호
• /
• pp.1257-1265
• /
• 2022

Recognition of sign language is challenging due to the occlusion of hands, accuracy of hand gestures, and high computational costs. In recent years, deep learning techniques have made significant advances in this field. Although these methods are larger and more complex, they cannot manage long-term sequential data and lack the ability to capture useful information through efficient information processing with faster convergence. In order to overcome these challenges, we propose a word-level sign language recognition (SLR) system that combines a real-time human pose detection library with the minimized version of the gated recurrent unit (GRU) model. Each gate unit is optimized by discarding the depth-weighted reset gate in GRU cells and considering only current input. Furthermore, we use sigmoid rather than hyperbolic tangent activation in standard GRUs due to performance loss associated with the former in deeper networks. Experimental results demonstrate that our pose-based optimized GRU (Pose-OGRU) outperforms the standard GRU model in terms of prediction accuracy, convergency, and information processing capability.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2009년도 춘계학술대회 논문집
• /
• pp.607-608
• /
• 2009

• 자동차안전학회지
• /
• 제14권3호
• /
• pp.71-76
• /
• 2022

In autonomous driving cars, 3D pose estimation can be one of the effective methods to enhance safety control for OOP (Out of Position) passengers. There have been many studies on human pose estimation using a camera. Previous methods, however, have limitations in automotive applications. Due to unexplainable failures, CNN methods are unreliable, and other methods perform poorly. This paper proposes robust real-time multi-human 3D pose estimation architecture in vehicle using monocular RGB camera. Using particle filter, our approach integrates CNN 2D/3D pose measurements with available information in vehicle. Computer simulations were performed to confirm the accuracy and robustness of the proposed algorithm.

• 인터넷정보학회논문지
• /
• 제21권5호
• /
• pp.21-29
• /
• 2020

Accurate estimation of human pose relies on backbone method in which its role is to extract feature map. Up to dated, the method of backbone feature extraction is conducted by the plain convolutional neural networks named by CNN and the residual neural networks named by Resnet, both of which have various architectures and performances. The CNN family network such as VGG which is well-known as a multiple stacked hidden layers architecture of deep learning methods, is base and simple while Resnet which is a bottleneck layers architecture yields fewer parameters and outperform. They have achieved inspired results as a backbone network in human pose estimation. However, they were used then followed by different pose estimation networks named by pose parsing module. Therefore, in this paper, we present a comparison between the plain CNN family network (VGG) and bottleneck network (Resnet) as a backbone method in the same pose parsing module. We investigate their performances such as number of parameters, loss score, precision and recall. We experiment them in the bottom-up method of human pose estimation system by adapted the pose parsing module of openpose. Our experimental results show that the backbone method using VGG network outperforms the Resent network with fewer parameter, lower loss score and higher accuracy of precision and recall.

• 한국정보통신학회논문지
• /
• 제25권9호
• /
• pp.1183-1189
• /
• 2021

홈 트레이닝을 하는 사람들은 전문적인 대면 지도가 없기 때문에 잘못된 자세로 동작을 하여 신체에 무리가 올 수 있다. 본 연구에서는 자세 예측 모델과 다층 퍼셉트론을 이용하여 사용자의 자세를 교정 해주는 "영상 데이터 기반 동작 분류 및 자세 교정 시스템"을 제안한다. 자세 예측 모델로 뼈대 정보를 예측한 후 심층 신경망을 이용하여 어떤 운동 동작인지를 분류한 뒤, 올바른 관절의 각도를 알려주며 교정이 이루어진다. 이 과정에서 동작 분류 모델의 성능을 향상시키기 위해 연속적인 프레임들의 결과를 고려하는 투표 알고리즘을 적용하였다. 다층 퍼셉트론 기반 모델을 자세 분류 모델로 사용했을 때 0.9의 정확도를 가진다. 그리고 투표 알고리즘을 통해 분류 모델의 정확도는 0.93으로 향상된다.