• Smart Media Journal
• /
• v.11 no.1
• /
• pp.31-37
• /
• 2022

This paper proposes a study on applying signal processing and deep learning for sound recognition that detects sounds commonly heard in daily life (Screaming, Clapping, Crowd_clapping, Car_passing_by and Back_ground, etc.). In the proposed sound recognition, several techniques related to the spectrum of sound waves, augmentation of sound data, ensemble learning for various predictions, convolutional neural networks (CNN) deep learning, and two-dimensional (2-D) data are used for improving the recognition accuracy. The proposed sound recognition technology shows that it can accurately recognize various sounds through experiments.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.5
• /
• pp.1778-1797
• /
• 2021

Document image binarization is an important pre-processing step in document analysis and archiving. The state-of-the-art models for document image binarization are variants of encoder-decoder architectures, such as FCN (fully convolutional network) and U-Net. Despite their success, they still suffer from three limitations: (1) reduced feature map resolution due to consecutive strided pooling or convolutions, (2) multiple scales of target objects, and (3) reduced localization accuracy due to the built-in invariance of deep convolutional neural networks (DCNNs). To overcome these three challenges, we propose an improved semantic segmentation model, referred to as DP-LinkNet, which adopts the D-LinkNet architecture as its backbone, with the proposed hybrid dilated convolution (HDC) and spatial pyramid pooling (SPP) modules between the encoder and the decoder. Extensive experiments are conducted on recent document image binarization competition (DIBCO) and handwritten document image binarization competition (H-DIBCO) benchmark datasets. Results show that our proposed DP-LinkNet outperforms other state-of-the-art techniques by a large margin. Our implementation and the pre-trained models are available at https://github.com/beargolden/DP-LinkNet.

• Journal of Broadcast Engineering
• /
• v.26 no.4
• /
• pp.441-452
• /
• 2021

In this paper, we propose a method for generating luma signal mapping function to improve the coding efficiency of luma signal mapping methods in LMCS. In this paper, we propose a method to reflect the cognitive and perceptual features by multiplying the attention map of convolutional neural networks on local spatial variance used to reflect local features in the existing LMCS. To evaluate the performance of the proposed method, BD-rate is compared with VTM-12.0 using classes A1, A2, B, C and D of MPEG standard test sequences under AI (All Intra) conditions. As a result of experiments, the proposed method in this paper shows improvement in performance the average of -0.07% for luma components in terms of BD-rate performance compared to VTM-12.0 and encoding/decoding time is almost the same.

• Journal of Korea Multimedia Society
• /
• v.22 no.9
• /
• pp.1000-1010
• /
• 2019

Image semantic segmentation and dehazing are key tasks in the computer vision. In recent years, researches in both tasks have achieved substantial improvements in performance with the development of Convolutional Neural Network (CNN). However, most of the previous works for semantic segmentation assume the images are captured in clear weather and show degraded performance under hazy images with low contrast and faded color. Meanwhile, dehazing aims to recover clear image given observed hazy image, which is an ill-posed problem and can be alleviated with additional information about the image. In this work, we propose a deep multi-task network for simultaneous semantic segmentation and dehazing. The proposed network takes single haze image as input and predicts dense semantic segmentation map and clear image. The visual information getting refined during the dehazing process can help the recognition task of semantic segmentation. On the other hand, semantic features obtained during the semantic segmentation process can provide cues for color priors for objects, which can help dehazing process. Experimental results demonstrate the effectiveness of the proposed multi-task approach, showing improved performance compared to the separate networks.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.11a
• /
• pp.62-65
• /
• 2018

신경망은 깊어질수록 gradient vanishing/exploding과 같은 네트워크가 불안정해지는 문제가 발생 한다. 잔차 블록을 이용하여 이러한 문제를 해결 할 수 있다. 본 논문에서는 영상 인식 분야에서 훌륭한 성능을 보여준 잔차 블록 기반의 깊은 합성곱 신경망을 통한 단일 영상 초해상도 복원 기법을 제안 한다. 제안한 알고리듬은 EDSR에 사용된 잔차 블록을 다양한 크기의 합성곱 연산을 통해 영상의 특징들을 다르게 분석하도록 수정하고 VDSR과 비슷한 수준의 복잡도로 구성하여 향상된 성능을 얻었다. 실험 결과, VDSR에 비해 PSNR이 최대 0.1dB까지 증가했다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.605-607
• /
• 2023

Recent research has highlighted the effectiveness of Deep Learning (DL) techniques in automating the detection of lung sound anomalies. However, the available lung sound datasets often suffer from limitations in both size and balance, prompting DL methods to employ data preprocessing such as augmentation and transfer learning techniques. These strategies, while valuable, contribute to the increased complexity of DL models and necessitate substantial training memory. In this study, we proposed a streamlined and lightweight DL method but effectively detects lung sound anomalies from small and imbalanced dataset. The utilization of 1D dilated convolutional neural networks enhances sensitivity to lung sound anomalies by efficiently capturing deep temporal features and small variations. We conducted a comprehensive evaluation of the ICBHI dataset and achieved a notable improvement over state-of-the-art results, increasing the average score of sensitivity and specificity metrics by 2.7%.

• Computers and Concrete
• /
• v.32 no.5
• /
• pp.475-486
• /
• 2023

Nondestructive evaluation (NDE) is an important task of civil engineering structure monitoring and inspection, but minor damage such as small cracks in local structure is difficult to observe. If cracks continued expansion may cause partial or even overall damage to the structure. Therefore, monitoring and detecting the structure in the early stage of crack propagation is important. The crack detection technology based on machine vision has been widely studied, but there are still some problems such as bad recognition effect for small cracks. In this paper, we proposed a deep learning method based on sweep signals to evaluate concrete surface crack with a width less than 1 mm. Two convolutional neural networks (CNNs) are used to analyze the one-dimensional (1D) frequency sweep signal and the two-dimensional (2D) time-frequency image, respectively, and the probability value of average damage (ADPV) is proposed to evaluate the minor damage of structural. Finally, we use the standard deviation of energy ratio change (ERVSD) and infrared thermography (IRT) to compare with ADPV to verify the effectiveness of the method proposed in this paper. The experiment results show that the method proposed in this paper can effectively predict whether the concrete surface is damaged and the severity of damage.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.3
• /
• pp.99-104
• /
• 2022

Convolutional neural network (CNN) generally uses two-stream architecture RGB and optical flow stream for its action recognition function. RGB frames stream display appearance and optical flow stream interprets its action. However, the standard method of using optical flow is costly in its computational time and latency associated with increased action recognition. The purpose of the study was to evaluate a novel way to create a two sub-networks in neural networks. The optical flow sub-network was assigned as a teacher and the RGB frames as a student. In the training stage, the optical flow sub-network extracts features through the teacher sub-network and transmits the information to student sub-network for baseline training. In the test stage, only student sub-network was operational with decreased in latency without computing optical flow. Experimental results shows that our network fed only by RGB stream gets a competitive accuracy of 54.5% on HMDB51, which is 1.5 times better than that on R3D-18.

• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.70-78
• /
• 2021

Estimating 3D information from a single image is one of the essential problems in numerous applications. Since a 2D image inherently might originate from an infinite number of different 3D scenes, thus 3D reconstruction from a single image is notoriously challenging. This challenge has been overcame by the advent of recent deep convolutional neural networks (CNNs), by modeling the mapping function between 2D image and 3D information. However, to train such deep CNNs, a massive training data is demanded, but such data is difficult to achieve or even impossible to build. Recent trends thus aim to present deep learning techniques that can be trained in a weakly-supervised manner, with a meta-data without relying on the ground-truth depth data. In this article, we introduce recent developments of weakly-supervised deep learning technique, especially categorized as scene 3D reconstruction and object 3D reconstruction, and discuss limitations and further directions.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.2
• /
• pp.27-37
• /
• 2021

In this paper, we propose Automatic detection system of underground pipe which automatically detects underground pipe to help experts. Actual location of underground pipe does not match with blueprint due to various factors such as ground changes over time, construction discrepancies, etc. So, various accidents occur during excavation or just by ageing. Locating underground utilities is done through GPR exploration to prevent these accidents but there are shortage of experts, because GPR data is enormous and takes long time to analyze. In this paper, To analyze 3D GPR data automatically, we use 3D image segmentation, one of deep learning technique, and propose proper data generation algorithm. We also propose data augmentation technique and pre-processing module that are adequate to GPR data. In experiment results, we found the possibility for pipe analysis using image segmentation through our system recorded the performance of F1 score 40.4%.