• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.8
• /
• pp.325-330
• /
• 2022

This paper conducts the collection and experiment of datasets for deep learning research on sign language such as sign language recognition, sign language translation, and sign language segmentation for Korean sign language. There exist difficulties for deep learning research of sign language. First, it is difficult to recognize sign languages since they contain multiple modalities including hand movements, hand directions, and facial expressions. Second, it is the absence of training data to conduct deep learning research. Currently, KETI dataset is the only known dataset for Korean sign language for deep learning. Sign language datasets for deep learning research are classified into two categories: Isolated sign language and Continuous sign language. Although several foreign sign language datasets have been collected over time. they are also insufficient for deep learning research of sign language. Therefore, we attempted to collect a large-scale Korean sign language dataset and evaluate it using a baseline model named TSPNet which has the performance of SOTA in the field of sign language translation. The collected dataset consists of a total of 11,402 image and text. Our experimental result with the baseline model using the dataset shows BLEU-4 score 3.63, which would be used as a basic performance of a baseline model for Korean sign language dataset. We hope that our experience of collecting Korean sign language dataset helps facilitate further research directions on Korean sign language.

• International journal of advanced smart convergence
• /
• v.6 no.4
• /
• pp.67-72
• /
• 2017

Deep learning neural network becomes very popular nowadays due to the reason that it can learn a very complex dataset such as the image dataset. Although deep learning neural network can produce high accuracy on the image dataset, it needs a lot of time to reach the convergence stage. To solve the issue, we have proposed a scaling method to improve the neural network to achieve the convergence stage in a shorter time than the original method. From the result, we can observe that our algorithm has higher performance than the other previous work.

• Journal of Korea Multimedia Society
• /
• v.22 no.10
• /
• pp.1133-1141
• /
• 2019

Most of the previous studies for bark recognition have focused on the extraction of LBP-like statistical features. Deep learning approach was not well studied because of the difficulty of acquiring large volume of bark image dataset. To overcome the bark dataset problem, this study utilizes the MobileNet which was trained with the ImageNet dataset. This study proposes two approaches. One is to extract features by the pixel-wise convolution and classify the features with SVM. The other is to tune the weights of the MobileNet by flexibly freezing layers. The experimental results with two public bark datasets, BarkTex and Trunk12, show that the proposed methods are effective in bark recognition. Especially the results of the flexible tunning method outperform state-of-the-art methods. In addition, it can be applied to mobile devices because the MobileNet is compact compared to other deep learning models.

• Korean Journal of Remote Sensing
• /
• v.37 no.6_1
• /
• pp.1685-1695
• /
• 2021

To generate a deep learning model with high performance, a large training dataset should be required. However, it requires a lot of time and cost to generate a large training dataset in remote sensing. Therefore, the importance of transfer learning of deep learning model using a small dataset have been increased. In this paper, we performed transfer learning of trained model based on open datasets by using orthoimages and digital maps to detect changes of building objects in multitemporal orthoimages. For this, an initial training was performed on open dataset for change detection through the HRNet-v2 model, and transfer learning was performed on dataset by orthoimages and digital maps. To analyze the effect of transfer learning, change detection results of various deep learning models including deep learning model by transfer learning were evaluated at two test sites. In the experiments, results by transfer learning represented best accuracy, compared to those by other deep learning models. Therefore, it was confirmed that the problem of insufficient training dataset could be solved by using transfer learning, and the change detection algorithm could be effectively applied to various remote sensed imagery.

• Imaging Science in Dentistry
• /
• v.52 no.2
• /
• pp.219-224
• /
• 2022

Purpose: This study aimed to evaluate the performance of transfer learning in a deep convolutional neural network for classifying implant fixtures. Materials and Methods: Periapical radiographs of implant fixtures obtained using the Superline (Dentium Co. Ltd., Seoul, Korea), TS III(Osstem Implant Co. Ltd., Seoul, Korea), and Bone Level Implant(Institut Straumann AG, Basel, Switzerland) systems were selected from patients who underwent dental implant treatment. All 355 implant fixtures comprised the total dataset and were annotated with the name of the system. The total dataset was split into a training dataset and a test dataset at a ratio of 8 to 2, respectively. YOLOv3 (You Only Look Once version 3, available at https://pjreddie.com/darknet/yolo/), a deep convolutional neural network that has been pretrained with a large image dataset of objects, was used to train the model to classify fixtures in periapical images, in a process called transfer learning. This network was trained with the training dataset for 100, 200, and 300 epochs. Using the test dataset, the performance of the network was evaluated in terms of sensitivity, specificity, and accuracy. Results: When YOLOv3 was trained for 200 epochs, the sensitivity, specificity, accuracy, and confidence score were the highest for all systems, with overall results of 94.4%, 97.9%, 96.7%, and 0.75, respectively. The network showed the best performance in classifying Bone Level Implant fixtures, with 100.0% sensitivity, specificity, and accuracy. Conclusion: Through transfer learning, high performance could be achieved with YOLOv3, even using a small amount of data.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.01a
• /
• pp.123-124
• /
• 2019

This paper proposes an improved deep learning method based on small data sets for animal image classification. Firstly, we use a CNN to build a training model for small data sets, and use data augmentation to expand the data samples of the training set. Secondly, using the pre-trained network on large-scale datasets, such as VGG16, the bottleneck features in the small dataset are extracted and to be stored in two NumPy files as new training datasets and test datasets. Finally, training a fully connected network with the new datasets. In this paper, we use Kaggle famous Dogs vs Cats dataset as the experimental dataset, which is a two-category classification dataset.

• International Journal of Computer Science & Network Security
• /
• v.23 no.11
• /
• pp.83-92
• /
• 2023

The majority of product users rely on the reviews that are posted on the appropriate website. Both users and the product's manufacturer could benefit from these reviews. Daily, thousands of reviews are submitted; how is it possible to read them all? Sentiment analysis has become a critical field of research as posting reviews become more and more common. Machine learning techniques that are supervised, unsupervised, and semi-supervised have worked very hard to harvest this data. The complicated and technological area of feature engineering falls within machine learning. Using deep learning, this tedious process may be completed automatically. Numerous studies have been conducted on deep learning models like LSTM, CNN, RNN, and GRU. Each model has employed a certain type of data, such as CNN for pictures and LSTM for language translation, etc. According to experimental results utilizing a publicly accessible dataset with reviews for all of the models, both positive and negative, and CNN, the best model for the dataset was identified in comparison to the other models, with an accuracy rate of 81%.

• Korean Journal of Remote Sensing
• /
• v.39 no.3
• /
• pp.269-282
• /
• 2023

The purpose of this study was to construct land cover maps using a deep learning model and to select the optimal deep learning model for land cover classification by adjusting the dataset such as input image size and Stride application. Two types of deep learning models, the U-net model and the DeeplabV3+ model with an Encoder-Decoder network, were utilized. Also, the combination of the two deep learning models, which is an Ensemble model, was used in this study. The dataset utilized RapidEye satellite images as input images and the label images used Raster images based on the six categories of the land use of Intergovernmental Panel on Climate Change as true value. This study focused on the problem of the quality improvement of the dataset to enhance the accuracy of deep learning model and constructed twelve land cover maps using the combination of three deep learning models (U-net, DeeplabV3+, and Ensemble), two input image sizes (64 × 64 pixel and 256 × 256 pixel), and two Stride application rates (50% and 100%). The evaluation of the accuracy of the label images and the deep learning-based land cover maps showed that the U-net and DeeplabV3+ models had high accuracy, with overall accuracy values of approximately 87.9% and 89.8%, and kappa coefficients of over 72%. In addition, applying the Ensemble and Stride to the deep learning models resulted in a maximum increase of approximately 3% in accuracy and an improvement in the issue of boundary inconsistency, which is a problem associated with Semantic Segmentation based deep learning models.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.178-179
• /
• 2018

Recently, methods of using machine learning in artificial intelligence such as autonomous navigation and speech recognition have been actively studied. Classical image processing methods such as classical boundary detection and pattern recognition have many limitations in order to recognize a specific object or area in a digital image. However, when a machine learning method such as deep-learning is used, Can be obtained. However, basically, a large amount of learning data must be secured for machine learning such as deep-learning. Therefore, it is difficult to apply the machine learning for area classification when the amount of data is very small, such as aerial photographs for environmental analysis. In this study, we apply a transfer-learning technique that can be used when the dataset size of the input image is small and the shape of the input image is not included in the category of the training dataset.

• Journal of information and communication convergence engineering
• /
• v.16 no.3
• /
• pp.173-178
• /
• 2018

Accurate classification of cloud images is a challenging task. Almost all the existing methods rely on hand-crafted feature extraction. Their limitation is low discriminative power. In the recent years, deep learning with convolution neural networks (CNNs), which can auto extract features, has achieved promising results in many computer vision and image understanding fields. However, deep learning approaches usually need large datasets. This paper proposes a deep learning approach for classification of cloud image patches on small datasets. First, we design a suitable deep learning model for small datasets using a CNN, and then we apply data augmentation and dropout regularization techniques to increase the generalization of the model. The experiments for the proposed approach were performed on SWIMCAT small dataset with k-fold cross-validation. The experimental results demonstrated perfect classification accuracy for most classes on every fold, and confirmed both the high accuracy and the robustness of the proposed model.