• 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 Korea Multimedia Society
• /
• v.24 no.11
• /
• pp.1492-1499
• /
• 2021

Skin prick test is widely used in diagnosing allergic sensitization to common inhalant or food allergens, in which positivities are manually determined by calculating the areas or mean diameters of wheals and erythemas provoked by allergens pricked into patients' skin. In this work, we propose a segmentation algorithm over U-Net, one of the FCN models of deep learning, to help us more objectively grasp the erythema boundaries. The performance of the model is analyzed by comparing the results of automatic segmentation of the test data to U-Net with the results of manual segmentation. As a result, the average Dice coefficient value was 94.93%, the average precision and sensitivity value was 95.19% and 95.24% respectively. We find that the proposed algorithm effectively discriminates the skin's erythema boundaries. We expect this algorithm to play an auxiliary role in skin prick test in real clinical trials in the future.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.105-116
• /
• 2022

Stay cables play an essential role in cable-stayed bridges. Severe vibrations and/or harsh environment may result in cable failures. Therefore, an efficient structural health monitoring (SHM) solution for cable damage detection is necessary. This study proposes a data-driven method for immediately detecting cable damage from measured cable forces by recognizing pattern transition from the intact condition when damage occurs. In the proposed method, pattern recognition for cable damage detection is realized by time series classification (TSC) using a deep learning (DL) model, namely, the long short term memory fully convolutional network (LSTM-FCN). First, a TSC classifier is trained and validated using the cable forces (or cable force ratios) collected from intact stay cables, setting the segmented data series as input and the cable (or cable pair) ID as class labels. Subsequently, the classifier is tested using the data collected under possible damaged conditions. Finally, the cable or cable pair corresponding to the least classification accuracy is recommended as the most probable damaged cable or cable pair. A case study using measured cable forces from an in-service cable-stayed bridge shows that the cable with damage can be correctly identified using the proposed DL-TSC method. Compared with existing cable damage detection methods in the literature, the DL-TSC method requires minor data preprocessing and feature engineering and thus enables fast and convenient early detection in real applications.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.7
• /
• pp.47-55
• /
• 2020

In this paper, we propose one stage multi-scale face detector based Fully Convolution Network using anchor free method. Recently almost all state-of-the-art face detectors which predict location of faces using anchor-based methods rely on pre-defined anchor boxes. However this face detectors need to hyper-parameters and additional computation in training. The key idea of the proposed method is to eliminate hyper-parameters and additional computation using anchor free method. To do this, we apply two ideas. First, by eliminating the pre-defined set of anchor boxes, we avoid the additional computation and hyper-parameters related to anchor boxes. Second, our detector predicts location of faces using multi-feature maps to reduce foreground/background imbalance issue. Through Quantitative evaluation, the performance of the proposed method is evaluated and analyzed. Experimental results on the FDDB dataset demonstrate the effective of our proposed method.

• Journal of Biomedical Engineering Research
• /
• v.41 no.1
• /
• pp.48-54
• /
• 2020

In this paper, we deal with a liver fibrosis classification problem using ultrasound B-mode images. Commonly representative methods for classifying the stages of liver fibrosis include liver biopsy and diagnosis based on ultrasound images. The overall liver shape and the smoothness and roughness of speckle pattern represented in ultrasound images are used for determining the fibrosis stages. Although the ultrasound image based classification is used frequently as an alternative or complementary method of the invasive biopsy, it also has the limitations that liver fibrosis stage decision depends on the image quality and the doctor's experience. With the rapid development of deep learning algorithms, several studies using deep learning methods have been carried out for automated liver fibrosis classification and showed superior performance of high accuracy. The performance of those deep learning methods depends closely on the amount of datasets. We propose an enhanced U-net architecture to maximize the classification accuracy with limited small amount of image datasets. U-net is well known as a neural network for fast and precise segmentation of medical images. We design it newly for the purpose of classifying liver fibrosis stages. In order to assess the performance of the proposed architecture, numerical experiments are conducted on a total of 118 ultrasound B-mode images acquired from 78 patients with liver fibrosis symptoms of F0~F4 stages. The experimental results support that the performance of the proposed architecture is much better compared to the transfer learning using the pre-trained model of VGGNet.

• The Journal of Korean Institute of Information Technology
• /
• v.16 no.12
• /
• pp.127-137
• /
• 2018

Parturition is an important event for farmers as it provides economic gains for the farms. Thus, the effective management of parturition is essential to farm management. In particular, the unit price of cattle is higher than other livestock and the productivity of cattle is closely associated to farm income. In addition, 42% of calving occurs in the nighttime so accurate parturition predictions are all the more important. In this paper, we propose a method that accurately predicts the calving date by applying core body temperature of cattle to deep learning. The body temperature of cattle can be measured without being influenced by the ambient environment by applying an ingestible bio-sensor in the cattle's rumen. By experiment on cattle, we confirmed this method to be more accurate for predicting calving dates than existing parturition prediction methods, showing an average of 3 hour 40 minute error. This proposed method is expected to reduce the economic damages of farms by accurately predicting calving times and assisting in successful parturitions.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.5
• /
• pp.55-67
• /
• 2022

Sensor data can provide fault diagnosis for equipment. However, the cause analysis for fault results of equipment is not often provided. In this study, we propose an explainable convolutional neural network framework for the sensor-based time series classification model. We used sensor-based time series dataset, acquired from vehicles equipped with sensors, and the Wafer dataset, acquired from manufacturing process. Moreover, we used Cycle Signal dataset, acquired from real world mechanical equipment, and for Data augmentation methods, scaling and jittering were used to train our deep learning models. In addition, our proposed classification models are convolutional neural network based models, FCN, 1D-CNN, and ResNet, to compare evaluations for each model. Our experimental results show that the ResNet provides promising results in the context of time series classification with accuracy and F1 Score reaching 95%, improved by 3% compared to the previous study. Furthermore, we propose XAI methods, Class Activation Map and Layer Visualization, to interpret the experiment result. XAI methods can visualize the time series interval that shows important factors for sensor data classification.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.223-223
• /
• 2023

토양수분은 지표면과 지하 영역 사이에 존재하는 수분 및 열에너지의 분배를 제어하거나, 토양 영양분, 식물 성장 및 미생물 활동과 같은 다양한 환경 과정에 영향을 미치는 핵심 구성요소이다. 토양수분은 생태수문학 및 생지화학적 역학, 저수지 관리, 가뭄 및 홍수의 경고, 토양 수분 변화에 따른 작물 수확량 등을 이해하는 데 매우 중요한 역할을 한다. 따라서, 토양 수분의 정확한 측정은 필수적이며, 이러한 필요성에 따라 중력 측정법, 장력 측정법, 전기 저항법 및 시간-주파수 영역반사측정법 등의 다양한 측정 방법들이 다년간 개발되어 사용되었다. 다만, 앞선 방법들은 철저한 실험을 통해 높은 정확성을 확보하였지만, 토양 교란이 발생하는 단점이 존재하며 실험 현장 토양의 물리적, 생물학적, 그리고 화학적 특성의 보존은 매우 어려운 한계점을 가지고 있다. 따라서, 이러한 단점을 극복하기 위해, 본 연구에서는 레일리파를 이용한 비접촉식 비교란 토양수분 센서 개발을 목표로 한다. 모래, 실트, 점토와 같은 세 가지 특징적인 토양 유형에 따른 파동을 측정하고, 측정된 파동으로부터 토양 수분을 추정하기 위해 기존에 개발된 시간-주파수 방법을 활용하여 토양수분을 함께 측정하였다. 비접촉 파동신호를 토양수분으로 변환하기 위하여, fully convolutional network을 개발하였다. 개발한 모델의 결과 검증은 RMSE(Root Mean Square Error)를 활용하여 검증하였으며, 모래, 실트, 점토에서 각각 0.0131, 0.0021, 0.0034 m³ m^-3으로 상대적으로 높은 정확성을 보였다. 즉, 본 연구에서 제시한 누출 레일리파를 사용한 비교란-비접촉 토양수분 측정 방법으로, 토양을 교란하지 않고 토양수분을 측정 할 수 있는 높은 가능성을 제시하였다.

• Journal of Internet Computing and Services
• /
• v.21 no.6
• /
• pp.23-31
• /
• 2020

This study was carried out to generate various images of railroad surfaces with random defects as training data to be better at the detection of defects. Defects on the surface of railroads are caused by various factors such as friction between track binding devices and adjacent tracks and can cause accidents such as broken rails, so railroad maintenance for defects is necessary. Therefore, various researches on defect detection and inspection using image processing or machine learning on railway surface images have been conducted to automate railroad inspection and to reduce railroad maintenance costs. In general, the performance of the image processing analysis method and machine learning technology is affected by the quantity and quality of data. For this reason, some researches require specific devices or vehicles to acquire images of the track surface at regular intervals to obtain a database of various railway surface images. On the contrary, in this study, in order to reduce and improve the operating cost of image acquisition, we constructed the 'Defective Railroad Surface Regeneration Model' by applying the methods presented in the related studies of the Generative Adversarial Network (GAN). Thus, we aimed to detect defects on railroad surface even without a dedicated database. This constructed model is designed to learn to generate the railroad surface combining the different railroad surface textures and the original surface, considering the ground truth of the railroad defects. The generated images of the railroad surface were used as training data in defect detection network, which is based on Fully Convolutional Network (FCN). To validate its performance, we clustered and divided the railroad data into three subsets, one subset as original railroad texture images and the remaining two subsets as another railroad surface texture images. In the first experiment, we used only original texture images for training sets in the defect detection model. And in the second experiment, we trained the generated images that were generated by combining the original images with a few railroad textures of the other images. Each defect detection model was evaluated in terms of 'intersection of union(IoU)' and F1-score measures with ground truths. As a result, the scores increased by about 10~15% when the generated images were used, compared to the case that only the original images were used. This proves that it is possible to detect defects by using the existing data and a few different texture images, even for the railroad surface images in which dedicated training database is not constructed.