• Journal of Platform Technology
• /
• v.6 no.1
• /
• pp.3-8
• /
• 2018

This paper proposes a technique for recognizing online handwritten cursive data obtained by tracing a motion trajectory while a user is in the 3D space based on a convolution neural network (CNN) algorithm. There is a difficulty in recognizing the virtual character input by the user in the 3D space because it includes both the character stroke and the movement stroke. In this paper, we divide syllable into consonant and vowel units by using labeling technique in addition to the result of localizing letter stroke and movement stroke in the previous study. The coordinate information of the separated consonants and vowels are converted into image data, and Korean handwriting recognition was performed using a convolutional neural network. After learning the neural network using 1,680 syllables written by five hand writers, the accuracy is calculated by using the new hand writers who did not participate in the writing of training data. The accuracy of phoneme-based recognition is 98.9% based on convolutional neural network. The proposed method has the advantage of drastically reducing learning data compared to syllable-based learning.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.6
• /
• pp.21-28
• /
• 2019

Recent years with the development of artificial intelligence and the success of the deep model, they have been deployed in all fields of computer vision. Action recognition, as an important branch of human perception and computer vision system research, has attracted more and more attention. Action recognition is a challenging task due to the special complexity of human movement, the same movement may exist between multiple individuals. The human action exists as a continuous image frame in the video, so action recognition requires more computational power than processing static images. And the simple use of the CNN network cannot achieve the desired results. Recently, the attention model has achieved good results in computer vision and natural language processing. In particular, for video action classification, after adding the attention model, it is more effective to focus on motion features and improve performance. It intuitively explains which part the model attends to when making a particular decision, which is very helpful in real applications. In this paper, we proposed a 3D dense convolutional network based on attention mechanism(ADD-Net), recognition of human motion behavior in the video.

• 대한방사선치료학회:학술대회논문집
• /
• 2005.06a
• /
• pp.13-17
• /
• 2005

Introduction : For stereotactic radiosurgery (SRS) of a tumor in the region whose movement due to respiration is significant, like Lung lower lobe, the gated therapy, which delivers radiation dose to the selected respiratory phases when tumor motion is small, was peformed using the Respiratory gating system and its clinical effectiveness was evaluated. Methode and Materials : For two SRS patients with a tumor in Lung lower lobe, a marker block (infrared reflector) was attached on the abdomen. While patient' respiratory cycle was monitored with Real-time Position Management (RPM, Varian, USA), 4D CT was performed (10 phases per a cycle). Phases in which tumor motion did not change rapidly were decided as treatment phases. The treatment volume was contoured on the CT images for selected treatment phases using maximum intensity projection (MIP) method. In order to verify setup reproducibility and positional variation, 4D CT was repeated. Result : Gross tumor volume (GTV) showed maximum movement in superior-inferior direction. For patient $\#$1, motion of GTV was reduced to 2.6 mm in treatment phases ($30\%\~60\%$), while that was 9.4 mm in full phases ($0\%\~90\%$) and for patient $\#$2, it was reduced to 2.3 mm in treatment phases ($30\%\~70\%$), while it was 11.7 mm in full phases ($0\%\~90\%$). When comparing two sets of CT images, setup errors in all the directions were within 3 mm. Conclusion : Since tumor motion was reduced less than 5 mm, the Respiratory gating system for SRS of Lung lower lobe is useful.

• Journal of Biomedical Engineering Research
• /
• v.28 no.6
• /
• pp.744-749
• /
• 2007

Kinematic and kinetic analysis has been performed for Soft Golf swings utilizing realistic three dimensional computer simulations based on three dimensional motion tracking data. Soft Golf is a newly developed recreational sport in South Korea aimed to become a safe and easy-to-learn sport for all ages. The advantage of Soft Golf stems from lighter weight of the club and much larger area of the sweet spot. This paper tries to look into kinematic and kinetic aspects of soft golf swings compared to regular golf swing and find the advantages of lighter Soft Golf clubs. For this purpose, swing motions of older aged participants were captured and kinematic analysis was performed for various kinematic parameters such as club head velocity, joint angular velocity, and joint range of motions as a pilot study. Kinetic analysis was performed by applying kinematic data to computer simulation models constructed from anthropometric database and the measurements from the participants. The simulations were solved using multi-body dynamics solver. Firstly, the kinematic parameters such as joint angles were obtained by solving inverse dynamics problem based on motion tracking data. Secondly, the kinetic parameters such as joint torques were obtained by solving control dynamics problem of making joint torque to follow pre-defined joint angle data. The results showed that mechanical loadings to major joints were reduced with lighter Soft Golf club.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.41 no.6
• /
• pp.91-99
• /
• 2004

This paper proposed two transcodiing methods, which maintain the same spatio-temporal resolution and reduce a spatial resolution, to convert a H.264 video bitstream into an MPEG-4 video bitstream. When the H.264 video bitstream is transformed into the MPEG-4 video bitstream, the conversions between H.264 block types and MPEG-4 block types are performed by minimizing distortion and the ${\times}4$ block-based motion vector mapping is performed. The proposed two transcoding methods run 4.14～5.1 times as fast as the cascaded transcoding methods in MPEG-4 encoder side, while the PSNR (peak-signal-to ratio) is slightly degrade with maximum 0.3dB.

• Radiation Oncology Journal
• /
• v.36 no.2
• /
• pp.122-128
• /
• 2018

Purpose: To investigate positional uncertainty and its correlation with clinical parameters in spine stereotactic body radiotherapy (SBRT) using thermoplastic mask (TM) immobilization. Materials and Methods: A total of 21 patients who underwent spine SBRT for cervical or upper thoracic spinal lesions were retrospectively analyzed. All patients were treated with image guidance using cone beam computed tomography (CBCT) and 4 degrees-of-freedom (DoF) positional correction. Initial, pre-treatment, and post-treatment CBCTs were analyzed. Setup error (SE), pre-treatment residual error (preRE), post-treatment residual error (postRE), intrafraction motion before treatment (IM1), and intrafraction motion during treatment (IM2) were determined from 6 DoF manual rigid registration. Results: The three-dimensional (3D) magnitudes of translational uncertainties (mean ${\pm}$ 2 standard deviation) were $3.7{\pm}3.5mm$ (SE), $0.9{\pm}0.9mm$ (preRE), $1.2{\pm}1.5mm$ (postRE), $1.4{\pm}2.4mm$ (IM1), and $0.9{\pm}1.0mm$ (IM2), and average angular differences were $1.1^{\circ}{\pm}1.2^{\circ}$ (SE), $0.9^{\circ}{\pm}1.1^{\circ}$ (preRE), $0.9^{\circ}{\pm}1.1^{\circ}$ (postRE), $0.6^{\circ}{\pm}0.9^{\circ}$ (IM1), and $0.5^{\circ}{\pm}0.5^{\circ}$ (IM2). The 3D magnitude of SE, preRE, postRE, IM1, and IM2 exceeded 2 mm in 18, 0, 3, 3, and 1 patients, respectively. No association were found between all positional uncertainties and body mass index, pain score, and treatment location (p > 0.05, Mann-Whitney test). There was a tendency of intrafraction motion to increase with overall treatment time; however, the correlation was not statistically significant (p > 0.05, Spearman rank correlation test). Conclusion: In spine SBRT using TM immobilization, CBCT and 4 DoF alignment correction, a minimum residual translational uncertainty was 2 mm. Shortening overall treatment time and 6 DoF positional correction may further reduce positional uncertainties.

• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.1-6
• /
• 2008

This paper proposes a game controller that provides user motion input and tactile feedback display, in addition to the traditional button-type input. The device utilizes both an accelerometer and an infrared camera in order to estimate 3D position and to recognize user motion. The information from the accelerometer and the camera are integrated for better performance. Various tactile sensations are presented using a voice-coil type vibrator. We apply the proposed controller to a motion-based game and validate its usability.

• Korean Journal of Applied Biomechanics
• /
• v.22 no.2
• /
• pp.173-181
• /
• 2012

The purpose of this study was to research on the movement of Fouette A La Second which was a type of turning movements on cheerleading. This research was conducted for helping cheerleaders to improve their overall skills. The three cheerleading national team members were participated in this research and the movements of Fouette A La Second were recorded with 6 digital motion master 60 video cameras, operating at a sampling frequency of 60 fields/sec. Six out of ten turning motion data were collected and analyzed with Kwon3D XP. The results were as follow: 1) The subject A's Releve motion was not executed precisely because of the COG's unstability. So she was required to improve the balancing ability. 2) The subject B could not execute the precise A La Second motion because of subject B's large hip angle. By tracing the projection of B's right toe on x-y plane, the subject made an elliptical orbit. Because B did not have a proper turning skills she needed to improve the muscle power and flexibility. She also needed to move quickly from Releve to Plie movement. 3) The subject C could not execute the Plie movement precisely, so she could not turn her body correctly around a certain spot. The subject C needed to decrease the knee angle at the Plie motion.

• Journal of the Korean Society of Physical Medicine
• /
• v.13 no.3
• /
• pp.113-120
• /
• 2018

PURPOSE: This study was conducted to investigate the characteristics of a specific functional shoe in terms of the range of motion (ROM) of ankle and foot joints during walking when compared to a standardized shoe. METHODS: Kinematic ROM data pertaining to ankle, tarsometatarsal, and metatarsophalangeal joints were collected from twenty-six healthy individuals during walking using a ten-camera motion analysis system. Kinematic ROM of each joint in three planes was obtained over ten walking trials consisting of two different shoe conditions. Visual3D motion analysis was finally used to coordinate the kinematic data. All kinematic ROM data were interpolated using a cubic spline algorithm and low-pass filtered with a cutoff frequency of 6 Hz for smoothing. RESULTS: The overall ROM of the ankle joint in the sagittal and coronal planes when wearing the specific functional shoe was significantly decreased in both ankles during walking when compared to wearing a standard shoe (p<.05). Significantly more flexibility was observed when wearing the specific functional shoe in the tarsometatarsal and metatarsophalangeal joints compared to a standard shoe (p<.05). CONCLUSION: Although clinical application of the specific functional shoe has shown clear positive effects on knee and ankle moments, the results of this study provide important background information regarding the kinematic mechanisms of these effects.

• Korean Journal of Computational Design and Engineering
• /
• v.15 no.6
• /
• pp.411-417
• /
• 2010

In this paper, the boom of a floating crane is modeled as a 3-dimensional elastic beam in order to analyze the dynamic response of the crane and its cargo. The boom is divided into more than two elements based on finite element formulation, and deformation of each element is expressed in terms of shape matrix and nodal coordinates. The equations of motion for the elastic boom consist of a mass matrix, a stiffness matrix, and a quadratic velocity vector that contains the gyroscopic and Coriolis forces. The size and complicity of the matrices increase in proportion with the number of elements. Therefore, it is not possible to derive the equations of motion explicitly for different number of elements. To overcome this difficulty, matrices for one 3-dimensional element are expressed with elementary sub-matrices. In particular, the quadratic velocity vector is derived as a product of a shape matrix and a 3-dimensional rotation matrix. By using the derived matrices, the equations of motion for the multi-element boom are automatically constructed. To verify the implementation of the elastic boom based on finite element formulation, we simulated a simple vibration of the elastic boom and compared the average deformation with the analytic solution. Finally, heave motion of the floating crane and surge motion of the cargo are presented as application examples of the elastic boom.