• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.05a
• /
• pp.97-100
• /
• 2016

Using the CNN(Convolutinal Neural Network), Deep Learning for variety of fields are being developed and these are showing significantly high level of performance at image recognition field. In this paper, we show the test accuracy which is learned by large-scale training data, over 5,000,000 of Korean characters. The architecture of CNN used in this paper is KCR(Korean Character Recognition)-AlexNet newly created based on AlexNet. KCR-AlexNet finally showed over 98% of test accuracy. The experimental data used in this paper is large-scale Korean character database PHD08 which has 2,187 samples for each Korean character and there are 2,350 Korean characters that makes total 5,139,450 sample data. Through this study, we show the excellence of architecture of KCR-AlexNet for learning PHD08.

• Journal of Information Processing Systems
• /
• v.14 no.1
• /
• pp.205-217
• /
• 2018

Deep learning using convolutional neural networks (CNNs) is being studied in various fields of image recognition and these studies show excellent performance. In this paper, we compare the performance of CNN architectures, KCR-AlexNet and KCR-GoogLeNet. The experimental data used in this paper is obtained from PHD08, a large-scale Korean character database. It has 2,187 samples of each Korean character with 2,350 Korean character classes for a total of 5,139,450 data samples. In the training results, KCR-AlexNet showed an accuracy of over 98% for the top-1 test and KCR-GoogLeNet showed an accuracy of over 99% for the top-1 test after the final training iteration. We made an additional Korean character dataset with fonts that were not in PHD08 to compare the classification success rate with commercial optical character recognition (OCR) programs and ensure the objectivity of the experiment. While the commercial OCR programs showed 66.95% to 83.16% classification success rates, KCR-AlexNet and KCR-GoogLeNet showed average classification success rates of 90.12% and 89.14%, respectively, which are higher than the commercial OCR programs' rates. Considering the time factor, KCR-AlexNet was faster than KCR-GoogLeNet when they were trained using PHD08; otherwise, KCR-GoogLeNet had a faster classification speed.

• Smart Media Journal
• /
• v.12 no.7
• /
• pp.52-58
• /
• 2023

Alzheimer's disease is one type of dementia, the symptoms can be treated by detecting the disease at its early stages. Recently, many computer-aided diagnosis using magnetic resonance image(MRI) have shown a good results in the classification of AD. Taken these MRI images and feed to Free surfer software to extra the features. In consideration, using T1-weighted images and classifying using the convolution neural network (CNN) model are proposed. In this paper, taking the subjects from ADNI of subcortical and cortical features of 190 subjects. Consider the study to reduce the complexity of the model by using the single layer in the Res-Net, VGG, and Alex Net. Multi-class classification is used to classify four different stages, CN, EMCI, LMCI, AD. The following experiment shows for respective classification Res-Net, VGG, and Alex Net with the best accuracy with VGG at 96%, Res-Net, GoogLeNet and Alex Net at 91%, 93% and 89% respectively.

• The Bulletin of The Korean Astronomical Society
• /
• v.42 no.2
• /
• pp.60.1-60.1
• /
• 2017

In this study we apply Convolution Neural Network(CNN) to solar flare occurrence prediction with various parameter options using the 00:00 UT MDI images from 1996 to 2010 (total 4962 images). We assume that only X, M and C class flares correspond to "flare occurrence" and the others to "non-flare". We have attempted to look for the best options for the models with two CNN pre-trained models (AlexNet and GoogLeNet), by modifying training images and changing hyper parameters. Our major results from this study are as follows. First, the flare occurrence predictions are relatively good with about 80 % accuracies. Second, both flare prediction models based on AlexNet and GoogLeNet have similar results but AlexNet is faster than GoogLeNet. Third, modifying the training images to reduce the projection effect is not effective. Fourth, skill scores of our flare occurrence model are mostly better than those of the previous models.

• Journal of Internet Computing and Services
• /
• v.18 no.5
• /
• pp.61-67
• /
• 2017

In this paper we introduce two experimental results from classification of Instagram images and some valuable lessons from them. We have tried some experiments for evaluating the competitive power of Convolutional Neural Network(CNN) in classification of real social network images such as Instagram images. We used AlexNet and ResNet, which showed the most outstanding capabilities in ImageNet Large Scale Visual Recognition Challenge(ILSVRC) 2012 and 2015, respectively. And we used 240 Instagram images and 12 pre-defined categories for classifying social network images. Also, we performed fine-tuning using Inception V3 model, and compared those results. In the results of four cases of AlexNet, ResNet, Inception V3 and fine-tuned Inception V3, the Top-1 error rates were 49.58%, 40.42%, 30.42%, and 5.00%. And the Top-5 error rates were 35.42%, 25.00%, 20.83%, and 0.00% respectively.

• Journal of Information Processing Systems
• /
• v.17 no.4
• /
• pp.754-771
• /
• 2021

In the task of continuous dimension emotion recognition, the parts that highlight the emotional expression are not the same in each mode, and the influences of different modes on the emotional state is also different. Therefore, this paper studies the fusion of the two most important modes in emotional recognition (voice and visual expression), and proposes a two-mode dual-modal emotion recognition method combined with the attention mechanism of the improved AlexNet network. After a simple preprocessing of the audio signal and the video signal, respectively, the first step is to use the prior knowledge to realize the extraction of audio characteristics. Then, facial expression features are extracted by the improved AlexNet network. Finally, the multimodal attention mechanism is used to fuse facial expression features and audio features, and the improved loss function is used to optimize the modal missing problem, so as to improve the robustness of the model and the performance of emotion recognition. The experimental results show that the concordance coefficient of the proposed model in the two dimensions of arousal and valence (concordance correlation coefficient) were 0.729 and 0.718, respectively, which are superior to several comparative algorithms.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.22 no.9
• /
• pp.1165-1171
• /
• 2018

CNN requires a large amount of computation and memory in the process of extracting the feature of the object. Also, It is trained from the network that the user has configured, and because the structure of the network is fixed, it can not be modified during training and it is also difficult to use it in a mobile device with low computing power. To solve these problems, we apply a pruning method to the pre-trained weight file to reduce computation and memory requirements. This method consists of three steps. First, all the weights of the pre-trained network file are retrieved for each layer. Second, take an absolute value for the weight of each layer and obtain the average. After setting the average to a threshold, remove the weight below the threshold. Finally, the network file applied the pruning method is re-trained. We experimented with LeNet-5 and AlexNet, achieved 31x on LeNet-5 and 12x on AlexNet.

• Journal of IKEEE
• /
• v.25 no.4
• /
• pp.664-671
• /
• 2021

This paper proposes a GPGPU structure that can reduce the number of operations and memory access by effectively applying a data reuse method to a convolutional neural network(CNN). Convolution is a two-dimensional operation using kernel and input data, and the operation is performed by sliding the kernel. In this case, a reuse method using an internal register is proposed instead of loading kernel from a cache memory until the convolution operation is completed. The serial operation method was applied to the convolution to increase the effect of data reuse by using the principle of GPGPU in which instructions are executed by the SIMT method. In this paper, for register-based data reuse, the kernel was fixed at 4×4 and GPGPU was designed considering the warp size and register bank to effectively support it. To verify the performance of the designed GPGPU on the CNN, we implemented it as an FPGA and then ran LeNet and measured the performance on AlexNet by comparison using TensorFlow. As a result of the measurement, 1-iteration learning speed based on AlexNet is 0.468sec and the inference speed is 0.135sec.

• Journal of Biomedical Engineering Research
• /
• v.39 no.6
• /
• pp.229-236
• /
• 2018

Liver cancer is one of the highest incidents in the world, and the mortality rate is the second most common disease after lung cancer. The purpose of this study is to evaluate the diagnostic ability of deep learning in the classification of malignant and benign tumors in CT images of patients with liver tumors. We also tried to identify the best data processing methods and deep learning models for classifying malignant and benign tumors in the liver. In this study, CT data were collected from 92 patients (benign liver tumors: 44, malignant liver tumors: 48) at the Gil Medical Center. The CT data of each patient were used for cross-sectional images of 3,024 liver tumors. In AlexNet and VggNet, the average of the overall accuracy at each image size was calculated: the average of the overall accuracy of the $200{\times}200$ image size is 69.58% (AlexNet), 69.4% (VggNet), $150{\times}150$ image size is 71.54%, 67%, $100{\times}100$ image size is 68.79%, 66.2%. In conclusion, the overall accuracy of each does not exceed 80%, so it does not have a high level of accuracy. In addition, the average accuracy in benign was 90.3% and the accuracy in malignant was 46.2%, which is a significant difference between benign and malignant. Also, the time it takes for AlexNet to learn is about 1.6 times faster than VggNet but statistically no different (p > 0.05). Since both models are less than 90% of the overall accuracy, more research and development are needed, such as learning the liver tumor data using a new model, or the process of pre-processing the data images in other methods. In the future, it will be useful to use specialists for image reading using deep learning.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.05a
• /
• pp.405-406
• /
• 2018

본 논문은 소표본 의료용 영상 분석의 정확도 향상을 위해 전이학습 모델을 feature extractor로 구축하여 학습시키는 방법을 연구하였으며 성능 평가를 위해 선학습모델로 AlexNet, ResNet, DenseNet을 사용하여 fine tuning 기법을 적용하였을 때와의 성능을 비교 분석하였다. 그 결과 실험에 사용된 3개의 모델에서 fine tuning 기법보다 향상된 정확도를 보임을 확인하였고, 또한 ImageNet으로 학습된 AlexNet, ResNet, DenseNet이 소표본 의료용 X-Ray 영상에 적용될 수 있음을 보였다.