The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
홍채 사진에서 영역을 정확하게 분할하는 것은 인식 시스템의 신뢰성에 중요한 영향을 미칩니다. 본 편지에서는 U-Net 기반의 엔드투엔드 심층신경망을 제시합니다. 조밀한 연결 블록을 사용하여 원래 컨벌루션 레이어를 대체하므로 피처 레이어의 재사용률을 효과적으로 향상시킬 수 있습니다. 제안된 방법은 업샘플링 과정(병합 계층)의 업샘플링 단계와 다운샘플링 단계의 동일 규모 특징 맵을 결합하기 위해 U-net의 스킵 연결을 사용합니다. 다운샘플링의 마지막 레이어에서는 확장된 컨볼루션을 사용합니다. 확장된 컨볼루션은 홍채 영역 위치 파악 정확도와 홍채 가장자리 픽셀 예측 정확도의 균형을 유지하여 네트워크 성능을 더욱 향상시킵니다. Casia v4 Interval 및 IITD 데이터 세트에서 실행되는 실험은 제안된 방법이 분할 성능을 향상시키는 것을 보여줍니다.
Chunhui GAO
Shanghai University
Guorui FENG
Shanghai University
Yanli REN
Shanghai University
Lizhuang LIU
Shanghai Advance Research Institute
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
부
Chunhui GAO, Guorui FENG, Yanli REN, Lizhuang LIU, "Iris Segmentation Based on Improved U-Net Network Model" in IEICE TRANSACTIONS on Fundamentals,
vol. E102-A, no. 8, pp. 982-985, August 2019, doi: 10.1587/transfun.E102.A.982.
Abstract: Accurate segmentation of the region in the iris picture has a crucial influence on the reliability of the recognition system. In this letter, we present an end to end deep neural network based on U-Net. It uses dense connection blocks to replace the original convolutional layer, which can effectively improve the reuse rate of the feature layer. The proposed method takes U-net's skip connections to combine the same-scale feature maps from the upsampling phase and the downsampling phase in the upsampling process (merge layer). In the last layer of downsampling, it uses dilated convolution. The dilated convolution balances the iris region localization accuracy and the iris edge pixel prediction accuracy, further improving network performance. The experiments running on the Casia v4 Interval and IITD datasets, show that the proposed method improves segmentation performance.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.E102.A.982/_p
부
@ARTICLE{e102-a_8_982,
author={Chunhui GAO, Guorui FENG, Yanli REN, Lizhuang LIU, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Iris Segmentation Based on Improved U-Net Network Model},
year={2019},
volume={E102-A},
number={8},
pages={982-985},
abstract={Accurate segmentation of the region in the iris picture has a crucial influence on the reliability of the recognition system. In this letter, we present an end to end deep neural network based on U-Net. It uses dense connection blocks to replace the original convolutional layer, which can effectively improve the reuse rate of the feature layer. The proposed method takes U-net's skip connections to combine the same-scale feature maps from the upsampling phase and the downsampling phase in the upsampling process (merge layer). In the last layer of downsampling, it uses dilated convolution. The dilated convolution balances the iris region localization accuracy and the iris edge pixel prediction accuracy, further improving network performance. The experiments running on the Casia v4 Interval and IITD datasets, show that the proposed method improves segmentation performance.},
keywords={},
doi={10.1587/transfun.E102.A.982},
ISSN={1745-1337},
month={August},}
부
TY - JOUR
TI - Iris Segmentation Based on Improved U-Net Network Model
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 982
EP - 985
AU - Chunhui GAO
AU - Guorui FENG
AU - Yanli REN
AU - Lizhuang LIU
PY - 2019
DO - 10.1587/transfun.E102.A.982
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E102-A
IS - 8
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - August 2019
AB - Accurate segmentation of the region in the iris picture has a crucial influence on the reliability of the recognition system. In this letter, we present an end to end deep neural network based on U-Net. It uses dense connection blocks to replace the original convolutional layer, which can effectively improve the reuse rate of the feature layer. The proposed method takes U-net's skip connections to combine the same-scale feature maps from the upsampling phase and the downsampling phase in the upsampling process (merge layer). In the last layer of downsampling, it uses dilated convolution. The dilated convolution balances the iris region localization accuracy and the iris edge pixel prediction accuracy, further improving network performance. The experiments running on the Casia v4 Interval and IITD datasets, show that the proposed method improves segmentation performance.
ER -