• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.3
• /
• pp.381-388
• /
• 2022

In instance segmentation, Mask-RCNN is mostly used as a base model. Increasing the performance of Mask-RCNN is meaningful because it affects the performance of the derived model. Mask-RCNN has a transform module for unifying size of input images. In this paper, to improve the Mask-RCNN, we apply deep-learning-based ASSR to the resizing part in the transform module and inject calculated scale information into the model using IM(Integration Module). The proposed IM improves instance segmentation performance by 2.5 AP higher than Mask-RCNN in the COCO dataset, and in the periment for optimizing the IM location, the best performance was shown when it was located in the 'Top' before FPN and backbone were combined. Therefore, the proposed method can improve the performance of models using Mask-RCNN as a base model.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.11
• /
• pp.491-500
• /
• 2021

In this study, we propose a method for detecting fall behavior using MS Kinect v2 RGBD Camera-based Human-Skeleton Keypoints and a 2-Stacked Bi-LSTM model. In previous studies, skeletal information was extracted from RGB images using a deep learning model such as OpenPose, and then recognition was performed using a recurrent neural network model such as LSTM and GRU. The proposed method receives skeletal information directly from the camera, extracts 2 time-series features of acceleration and distance, and then recognizes the fall behavior using the 2-Stacked Bi-LSTM model. The central joint was obtained for the major skeletons such as the shoulder, spine, and pelvis, and the movement acceleration and distance from the floor were proposed as features of the central joint. The extracted features were compared with models such as Stacked LSTM and Bi-LSTM, and improved detection performance compared to existing studies such as GRU and LSTM was demonstrated through experiments.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.4
• /
• pp.519-525
• /
• 2022

The face detection task of detecting a person's face in an image is used as a preprocess or core process in various image processing-based applications. The neural network models, which have recently been performing well with the development of deep learning, are dependent on 2D images, so if noise occurs in the image, such as poor camera quality or pool focus of the face, the face may not be detected properly. In this paper, we propose a face detection method that uses depth information together to reduce the dependence of 2D images. The proposed model was trained after generating and preprocessing depth information in advance using face detection dataset, and as a result, it was confirmed that the FRN model was 89.16%, which was about 1.2% better than the RetinaNet model, which showed 87.95%.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.1
• /
• pp.99-104
• /
• 2022

Artificial intelligence is getting a crucial part of our lives with its incredible benefits. Machines outperform humans in recognizing objects in images, particularly in classifying people into correct age and gender groups. In this respect, age and gender classification has been one of the hot topics among computer vision researchers in recent decades. Deployment of deep Convolutional Neural Network(: CNN) models achieved state-of-the-art performance. However, the most of CNN based architectures are very complex with several dozens of training parameters so they require much computation time and resources. For this reason, we propose a new CNN-based classification algorithm with significantly fewer training parameters and training time compared to the existing methods. Despite its less complexity, our model shows better accuracy of age and gender classification on the UTKFace dataset.

• Smart Media Journal
• /
• v.11 no.5
• /
• pp.48-55
• /
• 2022

Tomato crops are easy to expose to disease and spread in a short period of time, so late measures against disease are directly related to production and sales, which can cause damage. Therefore, there is a need for a service that enables early prevention by simply and accurately diagnosing tomato diseases in the field. In this paper, we construct a system that applies a deep learning-based model in which ImageNet transition is learned in advance to classify and serve nine classes of tomatoes for disease and normal cases. We use the input of MobileNet, ResNet, with a deep learning-based CNN structure that builds a lighter neural network using a composite product for the image set of leaves classifying tomato disease and normal from the Plant Village dataset. Through the learning of two proposed models, it is possible to provide fast and convenient services using MobileNet with high accuracy and learning speed.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.7
• /
• pp.307-314
• /
• 2022

Recently, researches using deep learning-based models are being actively conducted to replace statistical-based time series forecast techniques to predict electric power demand. The result of analyzing the researches shows that the performance of the LSTM-based prediction model is acceptable, but it is not sufficient for long-term regional-wide power demand prediction. In this paper, we propose a WaveNet deep learning model to predict electric power demand 24-hour-ahead with temperature data in order to achieve the prediction accuracy better than MAPE value of 2% which statistical-based time series forecast techniques can present. First of all, we illustrate a delated causal one-dimensional convolutional neural network architecture of WaveNet and the preprocessing mechanism of the input data of electric power demand and temperature. Second, we present the training process and walk forward validation with the modified WaveNet. The performance comparison results show that the prediction model with temperature data achieves MAPE value of 1.33%, which is better than MAPE Value (2.33%) of the same model without temperature data.

• Journal of IKEEE
• /
• v.26 no.3
• /
• pp.488-493
• /
• 2022

In manufacturing sites, bearing fault in eletrically driven motors cause the entire system to shut down. Stopping the operation of this environment causes huge losses in time and money. The reason of this bearing defects can be various factors such as wear due to continuous contact of rotating elements, excessive load addition, and operating environment. In this paper, a motor driving environment is created which is similar to the domestic manufacturing sites. In addition, based on the established environment, we propose a dataset for bearing fault detection by collecting changes in vibration characteristics that vary depending on normal and defective conditions. The sensor used to collect the vibration characteristics is Microphone G.R.A.S. 40PH-10. We used various machine learning models to build a prototype bearing fault detection system trained on the proposed dataset. As the result, based on the deep neural network model, it shows high accuracy performance of 92.3% in the time domain and 98.3% in the frequency domain.

• Journal of The Korean Astronomical Society
• /
• v.55 no.4
• /
• pp.111-121
• /
• 2022

Based on the light an exoplanet blocks from its host star as it passes in front of it during a transit, the mid-transit time can be determined. Periodic variations in mid-transit times can indicate another planet's gravitational influence. We investigate 83 transits of TrES-1 b as observed from 6-inch telescopes in the MicroObservatory robotic telescope network. The EXOTIC data reduction pipeline is used to process these transits, fit transit models to light curves, and calculate transit midpoints. This paper details the methodology for analyzing transit timing variations (TTVs) and using transit measurements to maintain ephemerides. The application of Lomb-Scargle period analysis for studying the plausibility of TTVs is explained. The analysis of the resultant TTVs from 46 transits from MicroObservatory and 47 transits from archival data in the Exoplanet Transit Database indicated the possible existence of other planets affecting the orbit of TrES-1 and improved the precision of the ephemeris by one order of magnitude. We now estimate the ephemeris to be (2 455 489.66026 BJD_TDB ± 0.00044 d) + (3.0300689 ± 0.0000007) d × epoch. This analysis also demonstrates the role of small telescopes in making precise midtransit time measurements, which can be used to help maintain ephemerides and perform TTV analysis. The maintenance of ephemerides allows for an increased ability to optimize telescope time on large ground-based telescopes and space telescope missions.

• Journal of Broadcast Engineering
• /
• v.27 no.5
• /
• pp.738-750
• /
• 2022

In this study, we reviewed and discussed whether auditory stimuli with short length can evoke emotion-related neurological responses. The findings implicate that if personalized sound tracks are provided to XR users based on machine learning or probability network models, user experiences in XR environment can be enhanced. We also investigated that the arousal-relaxed factor evoked by short auditory sound can make distinct patterns in functional connectivity characterized from background EEG signals. We found that coherence in the right hemisphere increases in sound-evoked arousal state, and vice versa in relaxed state. Our findings can be practically utilized in developing XR sound bio-feedback system which can provide preference sound to users for highly immersive XR experiences.

• Korean Journal of Remote Sensing
• /
• v.38 no.5_1
• /
• pp.559-570
• /
• 2022

When a satellite image has low spatial resolution, it is difficult to detect small objects. In this research, we aim to check the effect of super resolution on object detection. Super resolution is a software method that increases the resolution of an image. Unpaired super resolution network is used to improve Sentinel-2's spatial resolution from 10 m to 3.2 m. Faster-RCNN, RetinaNet, FCOS, and S²ANet were used to detect vessels in the Sentinel-2 images. We experimented the change in vessel detection performance when super resolution is applied. As a result, the Average Precision (AP) improved by at least 12.3% and up to 33.3% in the ship detection models trained with the super-resolution image. False positive and false negative cases also decreased. This implies that super resolution can be an important pre-processing step in object detection, and it is expected to greatly contribute to improving the accuracy of other image-based deep learning technologies along with object detection.