• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.3
• /
• pp.449-458
• /
• 2023

Artificial Intelligence is providing convenience in various fields using big data and deep learning technologies. However, deep learning technology is highly vulnerable to adversarial examples, which can cause misclassification of classification models. This study proposes a method to detect and purification various adversarial attacks using StarGAN. The proposed method trains a StarGAN model with added Categorical Entropy loss using adversarial examples generated by various attack methods to enable the Discriminator to detect adversarial examples and the Generator to purification them. Experimental results using the CIFAR-10 dataset showed an average detection performance of approximately 68.77%, an average purification performance of approximately 72.20%, and an average defense performance of approximately 93.11% derived from restoration and detection performance.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.11
• /
• pp.21-28
• /
• 2023

In manufacturing companies that focus on small-scale production of multiple product varieties, defective products are manually selected by workers rather than relying on automated inspection. Consequently, there is a higher risk of incorrect sorting due to variations in selection criteria based on the workers' experience and expertise, without consistent standards. Moreover, for non-standardized flexible objects with varying sizes and shapes, there can be even greater deviations in the selection criteria. To address these issues, this paper designs a quality inspection system using artificial intelligence-based unsupervised learning methods and conducts research by experimenting with accuracy using a dataset obtained from real manufacturing environments.

• The Journal of the Acoustical Society of Korea
• /
• v.43 no.1
• /
• pp.78-88
• /
• 2024

Ship-radiated noise received by passive sonar that can measure underwater noise can be identified and classified ship using Detection of Envelope Modulation on Noise (DEMON) analysis. However, in a low Signal-to-Noise Ratio (SNR) environment, it is difficult to analyze and identify the target frequency line containing ship information in the DEMONgram. In this paper, we conducted a study to extract target frequency lines using semantic segmentation among deep learning techniques for more accurate target identification in a low SNR environment. The semantic segmentation models U-Net, UNet++, and DeepLabv3+ were trained and evaluated using simulated DEMONgram data generated by changing SNR and fundamental frequency, and the DEMONgram prediction performance of DeepShip, a dataset of ship-radiated noise recordings on the strait of Georgia in Canada, was compared using the trained models. As a result of evaluating the trained model with the simulated DEMONgram, it was confirmed that U-Net had the highest performance and that it was possible to extract the target frequency line of the DEMONgram made by DeepShip to some extent.

• Journal of IKEEE
• /
• v.27 no.3
• /
• pp.313-318
• /
• 2023

This paper proposes a neural network applying fine-tuning as a way to improve the performance of Video Classification based on Vision Transformer. Recently, the need for real-time video image analysis based on deep learning has emerged. Due to the characteristics of the existing CNN model used in Image Classification, it is difficult to analyze the association of consecutive frames. We want to find and solve the optimal model by comparing and analyzing the Vision Transformer and Non-local neural network models with the Attention mechanism. In addition, we propose an optimal fine-tuning neural network model by applying various methods of fine-tuning as a transfer learning method. The experiment trained the model with the UCF101 dataset and then verified the performance of the model by applying a transfer learning method to the UTA-RLDD dataset.

• Journal of Korea Multimedia Society
• /
• v.23 no.8
• /
• pp.891-902
• /
• 2020

The Re-Identification(Re-ID) is one of the most popular researches in the field of computer vision due to a variety of applications. To achieve a high-level re-identification performance, recently other methods have developed the deep learning based networks that are specialized for only person or vehicle. However, most of the current methods are difficult to be used in real-world applications that require re-identification of both person and vehicle at the same time. To overcome this limitation, this paper proposes a deep neural network learning method that combines triplet and softmax loss to improve performance and re-identify people and vehicles simultaneously. It's possible to learn the detailed difference between the identities(IDs) by combining the softmax loss with the triplet loss. In addition, weights are devised to avoid bias in one-side loss when combining. We used Market-1501 and DukeMTMC-reID datasets, which are frequently used to evaluate person re-identification experiments. Moreover, the vehicle re-identification experiment was evaluated by using VeRi-776 and VehicleID datasets. Since the proposed method does not designed for a neural network specialized for a specific object, it can re-identify simultaneously both person and vehicle. To demonstrate this, an experiment was performed by using a person and vehicle re-identification dataset together.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.10
• /
• pp.411-420
• /
• 2019

Camera model identification has been a subject of steady study in the field of digital forensics. Among the increasingly sophisticated crimes, crimes such as illegal filming are taking up a high number of crimes because they are hard to detect as cameras become smaller. Therefore, technology that can specify which camera a particular image was taken on could be used as evidence to prove a criminal's suspicion when a criminal denies his or her criminal behavior. This paper proposes a deep learning model to identify the camera model used to acquire the image. The proposed model consists of four convolution layers and two fully connection layers, and a high pass filter is used as a filter for data pre-processing. To verify the performance of the proposed model, Dresden Image Database was used and the dataset was generated by applying the sequential partition method. To show the performance of the proposed model, it is compared with existing studies using 3 layers model or model with GLCM. The proposed model achieves 98% accuracy which is similar to that of the latest technology.

• Journal of Convergence for Information Technology
• /
• v.11 no.9
• /
• pp.28-34
• /
• 2021

In the case of object detection using deep learning, both accuracy and real-time are required. However, it is difficult to use a deep learning model that processes a large amount of data in a limited resource environment. To solve this problem, this paper proposes an object detection model for small embedded devices. Unlike the general detection model, the model size was minimized by using a structure in which the pre-trained feature extractor was removed. The structure of the model was designed by repeatedly stacking lightweight convolution blocks. In addition, the number of region proposals is greatly reduced to reduce detection overhead. The proposed model was trained and evaluated using the public dataset PASCAL VOC. For quantitative evaluation of the model, detection performance was measured with average precision used in the detection field. And the detection speed was measured in a Raspberry Pi similar to an actual embedded device. Through the experiment, we achieved improved accuracy and faster reasoning speed compared to the existing detection method.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.11
• /
• pp.1403-1410
• /
• 2020

Recently, as interest in non-face-to-face experiences and services increases, the demand for web video contents that can be easily consumed using mobile devices such as smartphones or tablets is rapidly increasing. To cope with these requirements, in this paper we propose a technique to efficiently produce video contents that can provide experience of visiting famous places (i.e., stage tour) in animation or movies. To this end, an image dataset was established by collecting images of stage areas using Google Maps and Google Street View APIs. Afterwards, a deep learning-based style transfer method to apply the unique style of animation videos to the collected street view images and generate the video contents from the style-transferred images was presented. Finally, we showed that the proposed method could produce more interesting stage-tour video contents through various experiments.

• Journal of IKEEE
• /
• v.24 no.4
• /
• pp.1122-1128
• /
• 2020

Bandwidth Extension refers to restoring and expanding a narrow band signal(NB) that is damaged or damaged in the encoding and decoding process due to the lack of channel capacity or the characteristics of the codec installed in the mobile communication device. It means converting to a wideband signal(WB). Bandwidth extension research mainly focuses on voice signals and converts high bands into frequency domains, such as SBR (Spectral Band Replication) and IGF (Intelligent Gap Filling), and restores disappeared or damaged high bands based on complex feature extraction processes. In this paper, we propose a model that outputs an bandwidth extended signal based on an autoencoder among deep learning models, using the residual connection of one-dimensional convolutional neural networks (CNN), the bandwidth is extended by inputting a time domain signal of a certain length without complicated pre-processing. In addition, it was confirmed that the damaged high band can be restored even by training on a dataset containing various types of sound sources including music that is not limited to the speech.

• Korean Journal of Remote Sensing
• /
• v.36 no.6_2
• /
• pp.1591-1604
• /
• 2020

This study is a study on classifying land cover by applying high-resolution satellite images to deep learning algorithms and verifying the performance of algorithms for each spatial object. For this, the Fully Convolutional Network-based algorithm was selected, and a dataset was constructed using Kompasat-3 satellite images, land cover maps, and forest maps. By applying the constructed data set to the algorithm, each optimal hyperparameter was calculated. Final classification was performed after hyperparameter optimization, and the overall accuracy of DeeplabV3+ was calculated the highest at 81.7%. However, when looking at the accuracy of each category, SegNet showed the best performance in roads and buildings, and U-Net showed the highest accuracy in hardwood trees and discussion items. In the case of Deeplab V3+, it performed better than the other two models in fields, facility cultivation, and grassland. Through the results, the limitations of applying one algorithm for land cover classification were confirmed, and if an appropriate algorithm for each spatial object is applied in the future, it is expected that high quality land cover classification results can be produced.