• Journal of Electrical Engineering and Technology
• /
• v.11 no.6
• /
• pp.1857-1862
• /
• 2016

Human activity recognition using depth information is an emerging and challenging technology in computer vision due to its considerable attention by many practical applications such as smart home/office system, personal health care and 3D video games. This paper presents a novel framework of 3D human body detection, tracking and recognition from depth video sequences using spatiotemporal features and modified HMM. To detect human silhouette, raw depth data is examined to extract human silhouette by considering spatial continuity and constraints of human motion information. While, frame differentiation is used to track human movements. Features extraction mechanism consists of spatial depth shape features and temporal joints features are used to improve classification performance. Both of these features are fused together to recognize different activities using the modified hidden Markov model (M-HMM). The proposed approach is evaluated on two challenging depth video datasets. Moreover, our system has significant abilities to handle subject's body parts rotation and body parts missing which provide major contributions in human activity recognition.

• Journal of Mechanical Science and Technology
• /
• v.18 no.12
• /
• pp.2095-2103
• /
• 2004

A macro-level CAPP system is proposed to plan the complicated mechanical prismatic parts efficiently. The system creates the efficient machining sequence of the features in a part by analyzing the feature information. Because the planning with the individual features is very complicated, feature groups are formed for effective planning using the nested relations of the features of a part, and special feature groups are determined for sequencing. The process plan is generated based on the sequences of the feature groups and features. When multiple machines are required, efficient machine assignment is performed. A series of heuristic rules are developed to accomplish it.

• ETRI Journal
• /
• v.42 no.3
• /
• pp.420-432
• /
• 2020

The reconstruction of archaeological fragments in 3D geometry is an important problem in pattern recognition and computer vision. Therefore, we implement an algorithm with the help of a 3D model to perform reconstruction from the real datasets using the slope features. This approach avoids the problem of gaps created through the loss of parts of the artifacts. Therefore, the aim of this study is to assemble the object without previous knowledge about the form of the original object. We utilize the edges of the fragments as an important feature in reconstructing the objects and apply multiple procedures to extract the 3D edge points. In order to assign the positions of the unknown parts that are supposed to match, the contour must be divided into four parts. Furthermore, to classify the fragments under reconstruction, we apply a backpropagation neural network. We test the algorithm on several models of ceramic fragments. It achieves highly accurate results in reconstructing the objects into their original forms, in spite of absent pieces.

• The Bulletin of The Korean Astronomical Society
• /
• v.43 no.1
• /
• pp.63.4-64
• /
• 2018

High resolution spectroscopy of the yellow symbiotic star AG Draconis is performed with the Canada-France-Hawaii Telescope to analyse the line profiles of Raman scattered O VI broad emission features at $6825{\AA}$ and $7082{\AA}$ with a view to investigating the wind accretion process from the mass losing giant to the white dwarf. These two spectral features are formed through inelastic scattering of O $VI{\lambda}{\lambda}32$ and 1038 with atomic hydrogen. We find that these features exhibit double-component profiles with red parts stronger than blue ones with the velocity separation of ~ 60 km s-1 in the O VI velocity space. Monte Carlo simulations for O VI line radiative transfer are performed by assuming that the O VI emission region constitutes a part of the accretion flow around the white dwarf and that Raman O VI features are formed in the neutral part of the slow stellar wind from the giant companion. The overall Raman O VI profiles are reasonably fit with an azimuthally asymmetric accretion flow and the mass loss rate ~ 4 ${\times}$ 10^{-7} M_sun yr^{-1}. We also find that additional bipolar neutral regions moving away with a speed ~ 70 km s^{-1} in the directions perpendicular to the orbital plane provide considerably improved fit to the red wing parts of Raman features.

• Journal of the Korean Society for Precision Engineering
• /
• v.7 no.3
• /
• pp.66-74
• /
• 1990

This paper presents a knowledge-based computer aided process planning system that automatically selects machine tools, machining operations and cutting tools and determines sequences of the machining operations for prismatic parts in die manufacturing. In the proposed system, parts are described by manufacturing features and grouped into part families based on the functions. Each part is repressented by a part frame which consists of basic data and manufacturing features. Knowledge for manufacturing is acquired from the domain expert and represented by frames. A decision model for selection of machine tools, machining operations and cutting tools and for determining sequences of the machining operations are developed by employing the Mealy machine in finite automata with output. The decision procedure and the order of priority which inputs manufacturing features into the Mealy machine are represented by rule for each part family. Backward chaining is used for the proposed system. The proposed system is implemented by using TURBO-PROLOG on the IBM PC/AT. A case study for the slide core is presented to show the function of the proposed system.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.25 no.4
• /
• pp.381-389
• /
• 2022

Detecting partially occluded objects is difficult due to the appearances and shapes of occluders are highly variable. These variabilities lead to challenges of localizing accurate bounding box or classifying objects with visible object parts. To address these problems, we propose a two-stage part-based attention approach for robust object detection under partial occlusion. First, our part attention network(PAN) captures the important object parts and then it is used to generate weighted object features. Based on the weighted features, the re-weighted object features are produced by our reinforced PAN(RPAN). Experiments are performed on our collected military vehicle dataset and synthetic occlusion dataset. Our method outperforms the baselines and demonstrates the robustness of detecting objects under partial occlusion.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.20 no.6
• /
• pp.1743-1752
• /
• 1996

In this paper, a precision inspection technique using CAD/CAI integration is proposed for the parts having very thin and sharp 3 dimensional curve features. The technique begings with feature reconstruction of turbine blades which have 3 dimensional combined feometry, such as splines, and thin circles. The alifnment procedures consistsb of two phases-rough and fine phases : rough phase alignment is based on the conventional 6 point5s probing on the clear cut surfacef, and fine phase alignment is based on the intial measurement on the 3 dimensional curved parts using an lterative measurement feed-back least sequares technique for alignment. Forf the analysis of profile tolerance of parts, the actual measured points are obtained by finding the closet points on the CAD geometry by the developed subdivision technique and the Tschebycheff norm is applied based on iterative fashion, giving accurate profile tolerance value. The developed inspection technique is applied to practical procedures of blade manufacturing and demonstrated high performance.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.17 no.12
• /
• pp.1380-1390
• /
• 1992

This paper proposes an on-line recognition system of Korean strokes using multi-layer neural network with tracing the stroke pattern. The system segments the stroke pattern into subpatterns, detects prominent stroke features in the subpatterns and integrates all the activation values of features in the related subpatterns. The activation values of the integrated stroke-specific features represent statistic characteristics of features and contributes for classifying the stroke pattern. Since the informations in Korean strokes are concentrated in the first and last parts of the strokes, the system extracts stroke-specific features in these parts attentatively and infers corner features using the sequential information of the extracted stroke-specific features in the first and last part of strokes the system is relatively simple in structure and rapid in on-line recognition of hand-written Korean strokes.

• Proceedings of the Safety Management and Science Conference
• /
• 2007.04a
• /
• pp.125-132
• /
• 2007

The purpose of this study is to provide foot classification on young men and women by factor analysis. The sample for this work was chosen from data which were collected and measured by Size Korea during two years ($2003{\sim}2004$). In order to analyze and compare features of the foot of young men and women, analysis was performed about 911 subjects (male: 467, female: 444) on 24 parts such as height (8 parts), width (5 parts), thickness (1 part), circumference (3 parts), length (3 parts) and angle (4 parts). The result of this study can be applied in manufacturing and design of shoes and socks. Also, it will enable us to have fruitful information on considerable items during manufacturing and design of shoes and socks.

• Journal of Korean Institute of Industrial Engineers
• /
• v.19 no.4
• /
• pp.23-40
• /
• 1993

In this paper, we developed a prototype part recognition system for FMS (Flexible Manufacturing System) by comparing the features extracted from CAD data and the features obtained from the image through a single camera, Good or distinct features with high recognition efficiency for a specific part are obtained automatically and intelligently according to the rules from CAD data. Since the selection of distinct features are done automatically without any human interference, the developed system seems to be quite robust. Fourther, since it uses only a few distinct features on-line among many features and all the time-consuming calculations are done off-line, it is possible to recognize each part quickly and accurately. In order to evaluate the performance and the effectiveness of the developed system, two example(L and T shaped parts) are tested and the results are reported.