• Journal of Convergence for Information Technology
• /
• v.11 no.10
• /
• pp.45-52
• /
• 2021

In this paper, we proposed an Atrous Residual U-Net (AR-UNet) to improve the segmentation accuracy of semantic segmentation method based on U-Net. The U-Net is mainly used in fields such as medical image analysis, autonomous vehicles, and remote sensing images. The conventional U-Net lacks extracted features due to the small number of convolution layers in the encoder part. The extracted features are essential for classifying object categories, and if they are insufficient, it causes a problem of lowering the segmentation accuracy. Therefore, to improve this problem, we proposed the AR-UNet using residual learning and ASPP in the encoder. Residual learning improves feature extraction ability and is effective in preventing feature loss and vanishing gradient problems caused by continuous convolutions. In addition, ASPP enables additional feature extraction without reducing the resolution of the feature map. Experiments verified the effectiveness of the AR-UNet with Cityscapes dataset. The experimental results showed that the AR-UNet showed improved segmentation results compared to the conventional U-Net. In this way, AR-UNet can contribute to the advancement of many applications where accuracy is important.

• ETRI Journal
• /
• v.44 no.6
• /
• pp.1034-1046
• /
• 2022

The learning-based multiview stereo (MVS) methods for three-dimensional (3D) reconstruction generally use 3D volumes for depth inference. The quality of the reconstructed depth maps and the corresponding point clouds is directly influenced by the spatial resolution of the 3D volume. Consequently, these methods produce point clouds with sparse local regions because of the lack of the memory required to encode a high volume of information. Here, we apply the atrous spatial pyramid pooling (ASPP) module in MVS methods to obtain dense feature maps with multiscale, long-range, contextual information using high receptive fields. For a given 3D volume with the same spatial resolution as that in the MVS methods, the dense feature maps from the ASPP module encoded with superior information can produce dense point clouds without a high memory footprint. Furthermore, we propose a 3D loss for training the MVS networks, which improves the predicted depth values by 24.44%. The ASPP module provides state-of-the-art qualitative results by constructing relatively dense point clouds, which improves the DTU MVS dataset benchmarks by 2.25% compared with those achieved in the previous MVS methods.

• Proceedings of the Korean Vacuum Society Conference
• /
• 1999.02a
• /
• pp.47-47
• /
• 1999

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.15-17
• /
• 2020

본 논문에서는 Convolutional Neural Network(CNN)를 이용한 위성 영상 변화탐지 알고리즘을 제안한다. 우선 EfficientNet 기반의 대칭 인코더-디코더 구조의 변화탐지 네트워크를 구성한다. 그리고 디코더 단에 ASPP 모듈을 추가하여 넓은 수용영역을 갖는 특징 정보를 통해 변화지도(change map)를 복원한다. 실험 결과를 통해 검출 성능 및 연산 효율성이 기존 기법보다 우수함을 보인다.

• Proceedings of the Korean Aquaculture Society Conference
• /
• 2003.10a
• /
• pp.19-19
• /
• 2003

This presentation reviewed the dietary vitamin C requirements in main marine finfish culture in Korea. In experiment I, an II-week feeding trial was conducted to study the effects of the different levels of dietary vitamin C on growth, tissue ascorbic acid concentrations and histopathological changes in parrot fish. Casein and gelatin based diets were formulated to contain 0, 60, 120, 240, 480 and 2000mg L-ascorbic acid (AA)kg diets on Ah equivalent basis in the form of L-ascorbyl-2-monophosphate (AMP), 60 and 240 mg AA/kg diet as L-ascorbic acid. Broken line analysis of weight gain indicated that the dietary vitamin C requirement of parrot fish is l18$\pm$12 mg AA/kg diet in the form of L-ascorbyl-2-monophosphate for maximum growth. In experiment II, a 12-week experiment was conducted to compare L-ascorbyl-2-monophosphate-Ca (AMP-Ca) with L-ascorbyl-2-monophosphate-Na/Ca (AMP-Na/Ca) for supplying the dietary vitamin C for juvenile Korean rockfish Sebastes schlegeli. Fish were fed one of 11 semi-purified diets containing equivalent of 0, 50, 100, 200, 400, and 800 mg ascorbic acid (AA)kg diet in the form of AMP-Ca or AMP-Na/Ca for 12 weeks. Broken line analysis of weight gain indicated that the dietary vitamin C requirement of Korean rockfish is 100 mg AA/kg diet in the form of AMP-Na/Ca, and 117 nag AA/kg diet in the form of AMP-Ca. In experiment III, a 12-week experiment was conducted to study the effects of different dietary levels of vitamin C, L-ascorbyl-2-polyphosphate (ASPP), on growth and tissue vitamin C concentrations in juvenile olive flounder. Fish were fed one of six semi-purified diets containing an equivalent of 0, 25, 50, 75, 150, or 1500 mg ascorbic acid (AA) kg 1 diet in the form of ASPP for 12 weeks. Based on broken line analyses for WG and PER, the optimum dietary levels of vitamin C were 91 and 93 mg AA/kg diet, respectively.

• Asian Pacific Journal of Cancer Prevention
• /
• v.13 no.5
• /
• pp.1727-1735
• /
• 2012

Regardless of advances in treatment modalities with the invention of newer therapies, breast cancer remains a major health problem with respect to its diagnosis, treatment and management. This female malignancy with its tremendous heterogeneous nature is linked to high incidence and mortality rates, especially in developing region of the world. It is the malignancy composed of distinct biological subtypes with diverse clinical, pathological, molecular and genetic features as well as different therapeutic responsiveness and outcomes. This inconsistency can be partially overcome by finding novel molecular markers with biological significance. In recent years, newer technologies help us to indentify distinct biomarkers and increase our understanding of the molecular basis of breast cancer. However, certain issues need to be resolved that limit the application of gene expression profiling to current clinical practice. Despite the complex nature of gene expression patterns of cDNAs in microarrays, there are some innovative regulatory molecules and functional pathways that allow us to predict breast cancer behavior in the clinic and provide new targets for breast cancer treatment. This review describes the landscape of different molecular markers with particular spotlight on vitamin D signaling pathway and apoptotic specific protein of p53 (ASPP) family members in breast cancer.

• Polymer(Korea)
• /
• v.25 no.1
• /
• pp.49-55
• /
• 2001

The bi-functional ion exchangers, SAPP-g-(AN/St) were synthesized with mixed vinyl monomers(acrylonitrile and styrene) onto PP fabric by the pre-irradiation grafting with E-beam and its subsequent amidoximination and sulfonation. The degree of grafting of PP-g-(AN/St) was increased with decreasing acrylonitrile composition in the mixed monomers. The water uptake of copolymers increased with decreasing in the amidoxime ratio in the copolymers and increased by sulfonation, but decreased by amidoximation. The $UO_2^{2+}$ adsorption capacity of SPP-g-St, APP-g-AN, and SAPP-g-(AN/St) were 12.4, 34.0, and 38.0 mg/g, respectively and the optimum adsorption time is about 50 hrs. As a result of uranium adsorption, the synthesized ion exchanger, which we obtained have also good affinity toward the adsorption or chelating with $UO_2^{2+}$ ions.

• The Journal of the Korea Contents Association
• /
• v.18 no.11
• /
• pp.447-454
• /
• 2018

This paper proposes a depth image generation algorithm of stereo images using a deep learning model composed of a CNN (convolutional neural network). The proposed algorithm consists of a feature extraction unit which extracts the main features of each parallax image and a depth learning unit which learns the parallax information using extracted features. First, the feature extraction unit extracts a feature map for each parallax image through the Xception module and the ASPP(Atrous spatial pyramid pooling) module, which are composed of 2D CNN layers. Then, the feature map for each parallax is accumulated in 3D form according to the time difference and the depth image is estimated after passing through the depth learning unit for learning the depth estimation weight through 3D CNN. The proposed algorithm estimates the depth of object region more accurately than other algorithms.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.1
• /
• pp.60-79
• /
• 2022

The U-Net architecture-based segmentation models attained remarkable performance in numerous medical image segmentation missions like skin lesion segmentation. Nevertheless, the resolution gradually decreases and the loss of spatial information increases with deeper network. The fusion of adjacent layers is not enough to make up for the lost spatial information, thus resulting in errors of segmentation boundary so as to decline the accuracy of segmentation. To tackle the issue, we propose a new deep learning-based segmentation model. In the decoding stage, the feature channels of each decoding unit are concatenated with all the feature channels of the upper coding unit. Which is done in order to ensure the segmentation effect by integrating spatial and semantic information, and promotes the robustness and generalization of our model by combining the atrous spatial pyramid pooling (ASPP) module and channel attention module (CAM). Extensive experiments on ISIC2016 and ISIC2017 common datasets proved that our model implements well and outperforms compared segmentation models for skin lesion segmentation.

• Journal of Microbiology and Biotechnology
• /
• v.24 no.5
• /
• pp.714-718
• /
• 2014

Apoptosis is the process of programmed cell death executed by specific proteases, the caspases, which mediate the cleavage of various vital proteins. Elucidating the consequences of this endoproteolytic cleavage is crucial to understanding cell death and other related biological processes. Although a number of possible roles for caspase-6 have been proposed, the identities and functions of proteins that interact with caspase-6 remain uncertain. In this study, we established a cell line expressing tandem affinity purification (TAP)-tagged caspase- 6 and then used LC-MS/MS proteomic analysis to analyze the caspase-6 interactome. Eight candidate caspase-6-interacting proteins were identified. Of these, five proteins (hnRNP-M, DHX38, ASPP2, MTA2, and UACA) were subsequently examined by co-immunoprecipitation for interactions with caspase-6. Thus, we identified two novel members of the caspase-6 interactome: hnRNP-M and MTA2.