• Proceedings of the KSRS Conference
• /
• 2008.10a
• /
• pp.386-389
• /
• 2008

It is important that the training stage of a supervised classification is designed to provide the spectral information. On the design of the training stage of a classification typically calls for the use of a large sample of randomly selected pure pixels in order to characterize the classes. Such guidance is generally made without regard to the specific nature of the application in-hand, including the classifier to be used. An approach to the training of a support vector machine (SVM) classifier that is the opposite of that generally promoted for training set design is suggested. This approach uses a small sample of mixed spectral responses drawn from purposefully selected locations (geographical boundaries) in training. A sample of such data should, however, be easier and cheaper to acquire than that suggested by traditional approaches. In this research, we evaluated them against traditional approaches with high-resolution satellite data. The results proved that it can be used small mixed pixels to derive a classification with similar accuracy using a large number of pure pixels. The approach can also reduce substantial costs in training data acquisition because the sampling locations used are commonly easy to observe.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.5
• /
• pp.479-487
• /
• 2021

In this paper, we propose an Sound Event Detection (SED) model using self-training based on a noisy student model. The proposed SED model consists of two stages. In the first stage, a mean-teacher model based on an Residual Convolutional Recurrent Neural Network (RCRNN) is constructed to provide target labels regarding weakly labeled or unlabeled data. In the second stage, a self-training-based noisy student model is constructed by applying different noise types. That is, feature noises, such as time-frequency shift, mixup, SpecAugment, and dropout-based model noise are used here. In addition, a semi-supervised loss function is applied to train the noisy student model, which acts as label noise injection. The performance of the proposed SED model is evaluated on the validation set of the Detection and Classification of Acoustic Scenes and Events (DCASE) 2020 Challenge Task 4. The experiments show that the single model and ensemble model of the proposed SED based on the noisy student model improve F1-score by 4.6 % and 3.4 % compared to the top-ranked model in DCASE 2020 challenge Task 4, respectively.

• International Journal of Contents
• /
• v.16 no.2
• /
• pp.8-18
• /
• 2020

Data is critical in deep learning but the scarcity of data often occurs in research, especially in the preparation of the paired training data. In this paper, document image binarization with unpaired data is studied by introducing adversarial learning, excluding the need for supervised or labeled datasets. However, the simple extension of the previous unpaired training to binarization inevitably leads to poor performance compared to paired data training. Thus, a new deep learning approach is proposed by introducing a multi-diversity of higher quality generated images. In this paper, a two-stage model is proposed that comprises the generative adversarial network (GAN) followed by the U-net network. In the first stage, the GAN uses the unpaired image data to create paired image data. With the second stage, the generated paired image data are passed through the U-net network for binarization. Thus, the trained U-net becomes the binarization model during the testing. The proposed model has been evaluated over the publicly available DIBCO dataset and it outperforms other techniques on unpaired training data. The paper shows the potential of using unpaired data for binarization, for the first time in the literature, which can be further improved to replace paired data training for binarization in the future.

• Archives of Plastic Surgery
• /
• v.41 no.3
• /
• pp.201-208
• /
• 2014

Microsurgery is a precise surgical skill that requires an extensive training period and the supervision of expert instructors. The classical training schemes in microsurgery have started with multiday experimental courses on the rat model. These courses have offered a low threat supervised high fidelity laboratory setting in which students can steadily and rapidly progress. This simulated environment allows students to make and recognise mistakes in microsurgery techniques and thus shifts any related risks of the early training period from the operating room to the lab. To achieve a high level of skill acquisition before beginning clinical practice, students are trained on a comprehensive set of exercises the rat model can uniquely provide, with progressive complexity as competency improves. This paper presents the utility of the classical rat model in three of the earliest microsurgery training centres and the new prospects that this versatile and expansive training model offers.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.12
• /
• pp.3889-3903
• /
• 2022

To address the problem of low detection accuracy due to training noise caused by mislabeling when Tri-training for network intrusion detection (NID), we propose a Tri-training algorithm based on cross entropy and K-nearest neighbors (TCK) for network intrusion detection. The proposed algorithm uses cross-entropy to replace the classification error rate to better identify the difference between the practical and predicted distributions of the model and reduce the prediction bias of mislabeled data to unlabeled data; K-nearest neighbors are used to remove the mislabeled data and reduce the number of mislabeled data. In order to verify the effectiveness of the algorithm proposed in this paper, experiments were conducted on 12 UCI datasets and NSL-KDD network intrusion datasets, and four indexes including accuracy, recall, F-measure and precision were used for comparison. The experimental results revealed that the TCK has superior performance than the conventional Tri-training algorithms and the Tri-training algorithms using only cross-entropy or K-nearest neighbor strategy.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.30 no.4
• /
• pp.27-43
• /
• 2005

This paper presents a new hybrid data mining technique using error pattern modeling to improve classification accuracy when the data type of a target variable is binary. The proposed method increases prediction accuracy by combining two different supervised learning methods. That is, the algorithm extracts a subset of training cases that are predicted inconsistently by both methods, and models error patterns from the cases. Based on the error pattern model, the Predictions of two different methods are merged to generate final prediction. The proposed method has been tested using practical 10 data sets. The analysis results show that the performance of proposed method is superior to the existing methods such as artificial neural networks and decision tree induction.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.13 no.1
• /
• pp.147-155
• /
• 2010

Landcover information of access-denied area was extracted from low-medium and high resolution satellite image. Training for supervised classification was performed to refer visually by landcover map which is made and distributed from The Ministry of Environment. The classification result was compared by relating data of FACC land classification system. As we rasterize digital military map with same pixel size of satellite classification, the accuracy test was performed by image to image method. In vegetation case, ancillary data such as NDVI and image for seasons are going to improve accuracy. FACC code of FDB need to recognize the properties which can be automated.

• International journal of advanced smart convergence
• /
• v.10 no.1
• /
• pp.216-224
• /
• 2021

Aim of this research was to classify forest types at Wando in Jeonnam Province and develop warm temperate forest management system with application of Remote Sensing and GIS. Another emphasis was given to the analysis of satellite images to compare forest type changes over 10 year periods from 2009 to 2019. We have accomplished this study by using ArcGIS Pro and ENVI. For this research, Landsat satellite images were obtained by means of terrestrial, airborne and satellite imagery. Based on the field survey data, all land uses and forest types were divided into 5 forest classes; Evergreen broad-leaved forest, Evergreen Coniferous forest, Deciduous broad-leaved forest, Mixed fores, and others. Supervised classification was carried out with a random forest classifier based on manually collected training polygons in ROI. Accuracy assessment of the different forest types and land-cover classifications was calculated based on the reference polygons. Comparison of forest changes over 10 year periods resulted in different vegetation biomass volumes, producing the loss of deciduous forests in 2019 probably due to the expansion of residential areas and rapid deforestation.

• Smart Media Journal
• /
• v.12 no.9
• /
• pp.45-59
• /
• 2023

Weakly supervised object localization (WSOL) is a task of localizing an object in an image using only image-level labels. Previous studies have followed the conventional class activation mapping (CAM) pipeline. However, we reveal the current CAM approach suffers from problems which cause original CAM could not capture the complete defects features. This work utilizes a convolutional neural network (CNN) pretrained on image-level labels to generate class activation maps in a multi-scale manner to highlight discriminative regions. Additionally, a vision transformer (ViT) pretrained was treated to produce multi-head attention maps as an auxiliary detector. By integrating the CNN-based CAMs and attention maps, our approach localizes defective regions without requiring bounding box or pixel-level supervision during training. We evaluate our approach on a dataset of apple images with only image-level labels of defect categories. Experiments demonstrate our proposed method aligns with several Object Detection models performance, hold a promise for improving localization.

• ETRI Journal
• /
• v.46 no.3
• /
• pp.513-525
• /
• 2024

Deep neural networks trained on labeled medical data face major challenges owing to the economic costs of data acquisition through expensive medical imaging devices, expert labor for data annotation, and large datasets to achieve optimal model performance. The heterogeneity of diseases, such as Alzheimer's disease, further complicates deep learning because the test cases may substantially differ from the training data, possibly increasing the rate of false positives. We propose a reconstruction-based self-supervised anomaly detection model to overcome these challenges. It has a dual-subnetwork encoder that enhances feature encoding augmented by skip connections to the decoder for improving the gradient flow. The novel encoder captures local and global features to improve image reconstruction. In addition, we introduce an entropy-based image conversion method. Extensive evaluations show that the proposed model outperforms benchmark models in anomaly detection and classification using an encoder. The supervised and unsupervised models show improved performances when trained with data preprocessed using the proposed image conversion method.