• The Sea:JOURNAL OF THE KOREAN SOCIETY OF OCEANOGRAPHY
• /
• v.10 no.1
• /
• pp.92-99
• /
• 2005

To discriminate the provenance of shelf sediments in the East China Sea, textural and elemental compositions along with strontium isotopic ratio ($^{87}Sr/^{86}Sr$) were analyzed and compared with the sediments originated from Chinese rivers. The sediments in the study area are composed of fine-grained mud with a mean grain size of $47\;{\phi}$ and their $CaCO_3$, contents range from 3.9 to 11.5% (average 7.6%). In the study area, the content of most metallic elements are strongly constrained by sediment grain size (quartz dilution effect) and that of biogenic material and, thereby, their spatial distribution seems not enough for understanding sediment provenance in the study area. The muddy sediments of the Yangtze river have much lower $^{87}Sr/^{86}Sr$ ratio ($0.71197{\sim}0.71720$) than the Yellow Sea shelf muddy sediments which are supposed to be originated from the Huanghe river ($0.72126{\sim}0.72498$), suggesting the distribution pattern of $^{87}Sr/^{86}Sr$ ratios as a new tracer to discriminate the provenance of shelf sediments in the study area. Different source rock compositions and weathering processes between both drainage basins may account for the differences in $^{87}Sr/^{86}Sr$ ratio. Although the ratios show wide range, from 0.71445 to 0.72184 with an average 0.71747 in the study area, they are close to the values of the Yangtze river sediments, suggesting that the sediments were mainly originated from the Yangtze river. The previous studies on the dispersal pattern of modern sediments and the physico-chemical properties of seawater in the Yellow and East China seas support the possibility that the fine-grained Yangtze river sediments can reach to the East China Sea shelf as well as to the southeastern Yellow Sea.

• Archives of design research
• /
• v.19 no.4 s.66
• /
• pp.133-144
• /
• 2006

According to the latest report by the Customer Protection Board, those who moved into newly constructed buildings are complaining about unidentified pains, asking for more careful selection of constructive materials for prevention of such potential problems. It is internationally recognized today that ecological materials can serve a significant factor for users' health, environmental protection and better industrial competitiveness. This study examined eco-design aspects of each interior material through web site search, in order to help customers learn about and capitalize on eco materials in a proper manner. As a result, 1. It turned out that the domestic industry are giving an impetus to releasing new eco items focusing on lower VOCs emission or addition of functional components as part of the marketing strategy. However, it is recommended that company understand significance of life cycle, and produce eco-concept materials. 2. The reliable standard for choosing the domestic material is EL, HB, GR marks. It is desirable to enhance recycling technologies and expand the sustainable consumption. customer class, since many recycled items are not developed. 3. The sourcing is a vulnerable part in terms of the concept of being environment-friendly material. Therefore, many manufacturers should design the easy knock-down products and produce the good items using recycled materials instead of new raw materials. Also solutions for making the energy from burning material should be studied. 4. The guidebook or manual with correct information about eco-materials is required to promote production and consumption with sustainable concept. 5. Many manufacturers are emphasizing ecological materials for customers, but some of them intended to disrupt customers' proper selection by promoting even unverified items to be environment-friendly.

• Korean Journal of Breeding Science
• /
• v.50 no.4
• /
• pp.406-414
• /
• 2018

This study was conducted to develop environmental risk assessments and biosafety guides for insect-resistant genetically modified rice in an LMO (Living Modified Organism) isolation field. In the LMO quarantine area of Kyungpook National University, the species diversities and population densities of non-target insects found on insect-resistant genetically modified rice (Bt-T), rice resistant to Cnaphalocrocis medinalis, and non-GM rice (Dongjin-byeo and Ilmi-byeo) were investigated. The Bt-T plants were, therefore, evaluated under field conditions to detect possible impacts on above ground insects and spiders. In 2016 and 2017, the study compared transgenic rice and two non-GM reference rice, namely Dongjin-byeo and Ilmi-byeo, at Gunwi. A total of 9,552 individuals from 51 families and 11 orders were collected from the LMO isolation field. From the three types of rice fields, a total of 3,042; 3,212; and 3,297 individuals from the Bt-T, Dongjin-byeo, and Ilmi-byeo were collected, respectively. There was no difference between the population densities of the non-target insect pests, natural enemies, and other insects on the Bt-T compared to non-GM rice. The data on insect species population densities were subjected to principal component analysis (PCA) without distinguishing between the three varieties, namely GM, non-GM, and reference cultivar, in all cultivation years. However, the PCA clearly separated the samples based on the cultivation years. These results suggest that insect species diversities and population densities during plant cultivation are determined by environmental factors (growing condition and seasons) rather than by genetic factors.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.