• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.8
• /
• pp.2787-2800
• /
• 2022

The aging society increases emergency situations of the elderly living alone and a variety of social crimes. In order to prevent them, techniques to detect emergency situations through voice are actively researched. This study proposes CutPaste-based anomaly detection model using multi-scale feature extraction in time series streaming data. In the proposed method, an audio file is converted into a spectrogram. In this way, it is possible to use an algorithm for image data, such as CNN. After that, mutli-scale feature extraction is applied. Three images drawn from Adaptive Pooling layer that has different-sized kernels are merged. In consideration of various types of anomaly, including point anomaly, contextual anomaly, and collective anomaly, the limitations of a conventional anomaly model are improved. Finally, CutPaste-based anomaly detection is conducted. Since the model is trained through self-supervised learning, it is possible to detect a diversity of emergency situations as anomaly without labeling. Therefore, the proposed model overcomes the limitations of a conventional model that classifies only labelled emergency situations. Also, the proposed model is evaluated to have better performance than a conventional anomaly detection model.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.27 no.12
• /
• pp.2110-2117
• /
• 2003

A physics-based material removal model in various scales is described and a feature scale simulation for a chemical mechanical polishing (CMP) process is performed in this work. Three different scales are considered in this model, i.e., abrasive particle scale, asperity scale and wafer scale. The abrasive particle and the asperity scales are combined together and then homogenized to result in force balance conditions to be satisfied in the wafer scale using an extended Greenwood-Williamson and Whitehouse-Archard statistical model that takes into consideration the joint distribution of asperity heights and asperity tip radii. The final computation is made to evaluate the material removal rate in wafer scale and a computer simulation is performed for detailed surface profile variations on a representative feature. The results show the dependence of the material removal rate on the joint distribution, applied external pressure, relative velocity, and other operating conditions and design parameters.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.4
• /
• pp.980-997
• /
• 2024

To improve the brightness of images and reveal hidden information in dark areas is the main objective of low-light image enhancement (LLIE). LLIE methods based on deep learning show good performance. However, there are some limitations to these methods, such as the complex network model requires highly configurable environments, and deficient enhancement of edge details leads to blurring of the target content. Single-scale feature extraction results in the insufficient recovery of the hidden content of the enhanced images. This paper proposed an edge detection-based multi-scale feature enhancement network for LLIE (EDMFEN). To reduce the loss of edge details in the enhanced images, an edge extraction module consisting of a Sobel operator is introduced to obtain edge information by computing gradients of images. In addition, a multi-scale feature enhancement module (MSFEM) consisting of multi-scale feature extraction block (MSFEB) and a spatial attention mechanism is proposed to thoroughly recover the hidden content of the enhanced images and obtain richer features. Since the fused features may contain some useless information, the MSFEB is introduced so as to obtain the image features with different perceptual fields. To use the multi-scale features more effectively, a spatial attention mechanism module is used to retain the key features and improve the model performance after fusing multi-scale features. Experimental results on two datasets and five baseline datasets show that EDMFEN has good performance when compared with the stateof-the-art LLIE methods.

• Journal of Korea Multimedia Society
• /
• v.21 no.12
• /
• pp.1467-1472
• /
• 2018

For the problem that the face image of surveillance video cannot be accurately identified due to the low resolution, this paper proposes a low resolution face recognition solution based on convolutional neural network model. Convolutional Neural Networks (CNN) model for multi-scale input The CNN model for multi-scale input is an improvement over the existing "two-step method" in which low-resolution images are up-sampled using a simple bi-cubic interpolation method. Then, the up sampled image and the high-resolution image are mixed as a model training sample. The CNN model learns the common feature space of the high- and low-resolution images, and then measures the feature similarity through the cosine distance. Finally, the recognition result is given. The experiments on the CMU PIE and Extended Yale B datasets show that the accuracy of the model is better than other comparison methods. Compared with the CMDA_BGE algorithm with the highest recognition rate, the accuracy rate is 2.5%~9.9%.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.10
• /
• pp.5161-5178
• /
• 2019

Correlation Filter was recently demonstrated to have good characteristics in the field of video object tracking. The advantages of Correlation Filter based trackers are reflected in the high accuracy and robustness it provides while maintaining a high speed. However, there are still some necessary improvements that should be made. First, most trackers cannot handle multi-scale problems. To solve this problem, our algorithm combines position estimation with scale estimation. The difference from the traditional method in regard to the scale estimation is that, the proposed method can track the scale of the object more quickly and effective. Additionally, in the feature extraction module, the feature representation of traditional algorithms is relatively simple, and furthermore, the tracking performance is easily affected in complex scenarios. In this paper, we design a novel and powerful feature that can significantly improve the tracking performance. Finally, traditional trackers often suffer from model drift, which is caused by occlusion and other complex scenarios. We introduce a relocation component to detect object at other locations such as the secondary peak of the response map. It partly alleviates the model drift problem.

• Korean Journal of Artificial Intelligence
• /
• v.11 no.3
• /
• pp.17-22
• /
• 2023

In this paper, we propose a black ice detection system using Convolutional Neural Networks (CNNs). Black ice poses a serious threat to road safety, particularly during winter conditions. To overcome this problem, we introduce a CNN-based architecture for real-time black ice detection with an encoder-decoder network, specifically designed for real-time black ice detection using thermal images. To train the network, we establish a specialized experimental platform to capture thermal images of various black ice formations on diverse road surfaces, including cement and asphalt. This enables us to curate a comprehensive dataset of thermal road black ice images for a training and evaluation purpose. Additionally, in order to enhance the accuracy of black ice detection, we propose a multi-scale dilation convolution feature fusion (MsDC-FF) technique. This proposed technique dynamically adjusts the dilation ratios based on the input image's resolution, improving the network's ability to capture fine-grained details. Experimental results demonstrate the superior performance of our proposed network model compared to conventional image segmentation models. Our model achieved an mIoU of 95.93%, while LinkNet achieved an mIoU of 95.39%. Therefore, it is concluded that the proposed model in this paper could offer a promising solution for real-time black ice detection, thereby enhancing road safety during winter conditions.

• Journal of Korea Multimedia Society
• /
• v.14 no.10
• /
• pp.1243-1251
• /
• 2011

This paper proposes a new vehicle model recognition using scale invariant feature transform to car's headlights image. Proposed vehicle model recognition raises the accuracy using "homogeneity" calculated from the distribution of features. In the experiment with 400 test images taken from 54 different vehicles, proposed method has 90% recognition rate and 16.45 homogeneity.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.5
• /
• pp.1431-1445
• /
• 2022

We construct a deep neural network model named ECGResNet. This model can diagnosis diseases based on 12-lead ECG data of eight common cardiovascular diseases with a high accuracy. We chose the 16 Blocks of ResNet50 as the main body of the model and added the Squeeze-and-Excitation module to learn the data information between channels adaptively. We modified the first convolutional layer of ResNet50 which has a convolutional kernel of 7 to a superposition of convolutional kernels of 8 and 16 as our feature extraction method. This way allows the model to focus on the overall trend of the ECG signal while also noticing subtle changes. The model further improves the accuracy of cardiovascular and cerebrovascular disease classification by using a fully connected layer that integrates factors such as gender and age. The ECGResNet model adds Dropout layers to both the residual block and SE module of ResNet50, further avoiding the phenomenon of model overfitting. The model was eventually trained using a five-fold cross-validation and Flooding training method, with an accuracy of 95% on the test set and an F1-score of 0.841.We design a new deep neural network, innovate a multi-scale feature extraction method, and apply the SE module to extract features of ECG data.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.4
• /
• pp.442-448
• /
• 2011

There are many research results about a self-localization technique of mobile robot. In this paper we present a self-localization technique based on the features of ceiling vision using a fisheye lens. The features obtained by SIFT(Scale Invariant Feature Transform) can be used to be matched between the previous image and the current image and then its optimal function is derived. The fisheye lens causes some distortion on its images naturally. So it must be calibrated by some algorithm. We here propose some methods for calibration of distorted images and design of a geometric fitness model. The proposed method is applied to laboratory and aile environment. We show its feasibility at some indoor environment.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.33B no.5
• /
• pp.60-72
• /
• 1996

In this paper, a enw model-base three-dimensional (3-D) object recognition mehtod using hopfield network is proposed. To minimize deformation of feature values on 3-D rotation, we select 3-D shape features and 3-D relational features which have rotational invariant characteristics. Then these feature values are normalized to have scale invariant characteristics, also. The input features are matched with model features by optimization process of hopjfield network in the form of two dimensional arrayed neurons. Experimental results on object classification and object matching with the 3-D rotated, scale changed, an dpartial oculued objects show good performance of proposed method.