• International Journal of Computer Science & Network Security
• /
• v.23 no.10
• /
• pp.37-43
• /
• 2023

The hospital situation, timing, and patient restrictions have become obstacles to an optimum therapy session. The crowdedness of the hospital might lead to a tight schedule and a shorter period of therapy. This condition might strike a post-stroke patient in a dilemma where they need regular treatment to recover their nervous system. In this work, we propose an in-house and uncomplex serious game system that can be used for physical therapy. The Kinect camera is used to capture the depth image stream of a human skeleton. Afterwards, the user might use their hand gesture to control the game. Voice recognition is deployed to ease them with play. Users must complete the given challenge to obtain a more significant outcome from this therapy system. Subjects will use their upper limb and hands to capture the 3D objects with different speeds and positions. The more substantial challenge, speed, and location will be increased and random. Each delegated entity will raise the scores. Afterwards, the scores will be further evaluated to correlate with therapy progress. Users are delighted with the system and eager to use it as their daily exercise. The experimental studies show a comparison between score and difficulty that represent characteristics of user and game. Users tend to quickly adapt to easy and medium levels, while high level requires better focus and proper synchronization between hand and eye to capture the 3D objects. The statistical analysis with a confidence rate(α:0.05) of the usability test shows that the proposed gaming is accessible, even without specialized training. It is not only for therapy but also for fitness because it can be used for body exercise. The result of the experiment is very satisfying. Most users enjoy and familiarize themselves quickly. The evaluation study demonstrates user satisfaction and perception during testing. Future work of the proposed serious game might involve haptic devices to stimulate their physical sensation.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.66 no.4
• /
• pp.27-39
• /
• 2024

Soil texture is determined by the proportions of sand, silt, and clay within the soil, which influence characteristics such as porosity, water retention capacity, electrical conductivity (EC), and pH. Traditional classification of soil texture requires significant sample preparation including oven drying to remove organic matter and moisture, a process that is both time-consuming and costly. This study aims to explore an alternative method by developing an AI model capable of predicting soil texture from images of pre-sorted soil samples using computer vision and deep learning technologies. Soil samples collected from agricultural fields were pre-processed using sieve analysis and the images of each sample were acquired in a controlled studio environment using a smartphone camera. Color distribution ratios based on RGB values of the images were analyzed using the OpenCV library in Python. A convolutional neural network (CNN) model, built on PyTorch, was enhanced using Digital Image Processing (DIP) techniques and then trained across nine distinct conditions to evaluate its robustness and accuracy. The model has achieved an accuracy of over 80% in classifying the images of pre-sorted soil samples, as validated by the components of the confusion matrix and measurements of the F1 score, demonstrating its potential to replace traditional experimental methods for soil texture classification. By utilizing an easily accessible tool, significant time and cost savings can be expected compared to traditional methods.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.3
• /
• pp.563-570
• /
• 2024

This paper proposes a method to recognize and track drivable lane areas to assist the driver. The main topic is designing a deep-based network that predicts drivable road areas using computer vision and deep learning technology based on images acquired in real time through a camera installed in the center of the windshield inside the vehicle. This study aims to develop a new model trained with data directly obtained from cameras using the YOLO algorithm. It is expected to play a role in assisting the driver's driving by visualizing the exact location of the vehicle on the actual road consistent with the actual image and displaying and tracking the drivable lane area. As a result of the experiment, it was possible to track the drivable road area in most cases, but in bad weather such as heavy rain at night, there were cases where lanes were not accurately recognized, so improvement in model performance is needed to solve this problem.

• Information Systems Review
• /
• v.20 no.3
• /
• pp.73-92
• /
• 2018

Many users are uploading their daily life activities on SNS and use hashtags to describe their postings. Hashtag has the advantage of letting users specify categories for their postings, however until now, the users has had to manually input the hashtags which has been very inconvenient for them. Therefore, in order to address this issue, this paper proposes a hashtag recommender system which recommends proper hashtags to users based on their uploaded images on SNS. The proposed system is designed using four analytic structures, which is composed of a camera information-based analysis, an address-based analysis, a location based CF analysis, and an image-based analysis. In order to check whether the proposed system is improved compared to the existing systems in terms of the hashtag recommendation function, we conducted an evaluation with 212 SNS users from fifteen countries. As a result of the evaluation process, the proposed system shows very high accuracy recommendation results compared to the existing hashtag recommender systems.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.3
• /
• pp.953-958
• /
• 2024

Recently, there has been growing interest in the Fourth Industrial Revolution, with a particular focus on the advancement of augmented reality (AR) devices. However, there is a shortage of AR content. Augmented reality operates through marker-based and markerless methods. The marker-based approach involves using a camera to capture images that serve as markers, enhancing them through AR principles. To address the scarcity of AR content and improve the quality of character goods, this study proposes integrating AR technology into character goods. The character industry is expanding each year, leading to a diverse range of character goods. Character acrylic stands, among these goods, leverage game, webtoon, and animation character IPs for sales. To enhance the design process, we utilized the character image as a marker, allowing for the creation of content that aligns with the characteristics of the character IP. We selected a webtoon character and developed AR content, incorporating features such as voice, speech bubbles, and an introduction to the webtoon, tailored to the webtoon's characteristics. This study demonstrates the potential of AR to present visual and auditory information, paving the way for a variety of products, including diverse content. We anticipate that utilizing this research will lead to the emergence of products encompassing various contents.

• Journal of the Korea Computer Graphics Society
• /
• v.30 no.3
• /
• pp.109-123
• /
• 2024

Research into vision-based end-to-end autonomous driving systems utilizing deep learning and reinforcement learning has been steadily increasing. These systems typically encode continuous and high-dimensional vehicle states, such as location, velocity, orientation, and sensor data, into latent features, which are then decoded into a vehicular control policy. The complexity of urban driving environments necessitates the use of state representation learning through networks like Variational Autoencoders (VAEs) or Convolutional Neural Networks (CNNs). This paper analyzes the impact of different image state encoding methods on reinforcement learning performance in autonomous driving. Experiments were conducted in the CARLA simulator using RGB images and semantically segmented images captured by the vehicle's front camera. These images were encoded using VAE and Vision Transformer (ViT) networks. The study examines how these networks influence the agents' learning outcomes and experimentally demonstrates the role of each state representation technique in enhancing the learning efficiency and decision- making capabilities of autonomous driving systems.

• International conference on construction engineering and project management
• /
• 2024.07a
• /
• pp.447-454
• /
• 2024

Construction site monitoring is pivotal for overseeing project progress to ensure that projects are completed as planned, within budget, and in compliance with applicable laws and safety standards. Additionally, it seeks to improve operational efficiency for better project execution. To achieve this, many researchers have utilized computer vision technologies to conduct automatic site monitoring and analyze the operational status of equipment. However, most existing studies estimate real-world 3D information (e.g., object tracking, work status analysis) based only on 2D pixel-based information of images. This approach presents a substantial challenge in the dynamic environments of construction sites, necessitating the manual recalibration of analytical rules and thresholds based on the specific placement and the field of view of cameras. To address these challenges, this study introduces a novel method for 3D visualization and status analysis of construction site objects using 3D reconstruction technology. This method enables the analysis of equipment's operational status by acquiring 3D spatial information of equipment from single-camera images, utilizing the Sam-Track model for object segmentation and the One-2-3-45 model for 3D reconstruction. The framework consists of three main processes: (i) single image-based 3D reconstruction, (ii) 3D visualization, and (iii) work status analysis. Experimental results from a construction site video demonstrated the method's feasibility and satisfactory performance, achieving high accuracy in status analysis for excavators (93.33%) and dump trucks (98.33%). This research provides a more consistent method for analyzing working status, making it suitable for practical field applications and offering new directions for research in vision-based 3D information analysis. Future studies will apply this method to longer videos and diverse construction sites, comparing its performance with existing 2D pixel-based methods.

• Nuclear Engineering and Technology
• /
• v.56 no.10
• /
• pp.4423-4436
• /
• 2024

The development and assessment of pharmaceutical dosage forms make considerable use of gamma-scintigraphy. Gamma scintigraphy is an imaging technique that is integrated with CT to assess and evaluate the targeting of drugs to various delivery sites, the impact of treatment, and the severity of the disease. A small number of radioisotopes were tagged with the delivery system and emitted radiation can be visualized by the gamma camera which forms a 2-D image displaying the tissue-specific distribution of radioactivity. The isotopes that are used widely include Technetium-99 m (99Tc), Iodine (131I), Fluorodeoxyglucose (18F-FDG), Fluoromisonidazole (18F-FMISO) and Gallium (Ga67), Indium (111In). This review mainly covers different applications of gamma scintigraphy for the assessment of drug targeting via different routes to different organs and their visualization by gamma scintigraphy. The review mainly focuses assessment of drug targeting in the tumor tissue, thyroid gland, brain, pulmonary pathway, skin deposition, detection of renal impairment as well as cardiac diseases, drug release from formulations, drug deposition in arthritis, drug retention in the scalp, and behavior of formulation when administered via intra-vaginal route. Various pre-clinical and clinical studies were included in the review that demonstrates the importance and future of gamma scintigraphy in sensing drug delivery.

• Investigative Magnetic Resonance Imaging
• /
• v.16 no.1
• /
• pp.55-66
• /
• 2012

Purpose : During brain MRI scanning, subject's head motion can adversely affect MRI images. To minimize MR image distortion by head movement, we developed an optical tracking system to detect the 3-D movement of subjects. Materials and Methods: The system consisted of 2 CCD cameras, two infrared illuminators, reflective sphere-type markers, and frame grabber with desktop PC. Using calibration which is the procedure to calculate intrinsic/extrinsic parameters of each camera and triangulation, the system was desiged to detect 3-D coordinates of subject's head movement. We evaluated the accuracy of 3-D position of reflective markers on both test board and the real MRI scans. Results: The stereo system computed the 3-D position of markers accurately for the test board and for the subject with glasses with attached optical reflective marker, required to make regular head motion during MRI scanning. This head motion tracking didn't affect the resulting MR images even in the environment varying magnetic gradient and several RF pulses. Conclusion: This system has an advantage to detect subject's head motion in real-time. Using the developed system, MRI operator is able to determine whether he/she should stop or intervene in MRI acquisition to prevent more image distortions.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.17 no.4
• /
• pp.307-314
• /
• 2011

This study proposed detection method of Marine debris using unmanned aerial photography. For unmanned aerial photography, a RC(Radio Control) helicopter which has good movability and economics was used. To a camera mounting, a gimbal equipment was attached to the bottom of the RC helicopter. The gimbal equipment is very useful because it is not seriously affected by vibration and rolling. In addition, we invented that digital image processing algorithm using Matlab program for detection of marine debris from photographs. Particularly, background subtraction in invented algorithm was applied. As a result, marine debris of a variety of forms from different sand states of coast were reliably detected. In the future, monitoring using proposed method was expected to contribute that the solution to representative problem of monitoring area selecting and estimate the total litter mass over the beach. Moreover, It is considered a greater application possibility to marine environmental observations.