TY - JOUR
T1 - Automated Skin Lesion Segmentation Via an Adaptive Dual Attention Module
AU - Wu, Huisi
AU - Pan, Junquan
AU - Li, Zhuoying
AU - Wen, Zhenkun
AU - Qin, Jing
N1 - Funding Information:
Manuscript received July 29, 2020; revised September 17, 2020 and September 22, 2020; accepted September 23, 2020. Date of publication September 28, 2020; date of current version December 29, 2020. This work was supported in part by the National Natural Science Foundation of China under Grant 61973221, in part by the Natural Science Foundation of Guangdong Province of China under Grant 2018A030313381 and Grant 2019A1515011165, in part by the Key Lab of Shenzhen Research Foundation of China under Grant 201707311550233, in part by the COVID-19 Prevention Project of Guangdong Province of China under Grant 2020KZDZX1174, in part by the Major Project of the New Generation of Artificial Intelligence of China under Grant 2018AAA0102900, and in part by the Hong Kong Research Grants Council of China under Grant PolyU 152035/17E and Grant 15205919. (Corresponding author: Huisi Wu.) Huisi Wu, Junquan Pan, Zhuoying Li, and Zhenkun Wen are with the College of Computer Science and Software Engineering, Shen-zhen University, Shenzhen 518060, China (e-mail: [email protected]; [email protected]; [email protected]; [email protected]).
Publisher Copyright:
© 1982-2012 IEEE.
PY - 2021/1
Y1 - 2021/1
N2 - We present a convolutional neural network (CNN) equipped with a novel and efficient adaptive dual attention module (ADAM) for automated skin lesion segmentation from dermoscopic images, which is an essential yet challenging step for the development of a computer-assisted skin disease diagnosis system. The proposed ADAM has three compelling characteristics. First, we integrate two global context modeling mechanisms into the ADAM, one aiming at capturing the boundary continuity of skin lesion by global average pooling while the other dealing with the shape irregularity by pixel-wise correlation. In this regard, our network, thanks to the proposed ADAM, is capable of extracting more comprehensive and discriminative features for recognizing the boundary of skin lesions. Second, the proposed ADAM supports multi-scale resolution fusion, and hence can capture multi-scale features to further improve the segmentation accuracy. Third, as we harness a spatial information weighting method in the proposed network, our method can reduce a lot of redundancies compared with traditional CNNs. The proposed network is implemented based on a dual encoder architecture, which is able to enlarge the receptive field without greatly increasing the network parameters. In addition, we assign different dilation rates to different ADAMs so that it can adaptively capture distinguishing features according to the size of a lesion. We extensively evaluate the proposed method on both ISBI2017 and ISIC2018 datasets and the experimental results demonstrate that, without using network ensemble schemes, our method is capable of achieving better segmentation performance than state-of-the-art deep learning models, particularly those equipped with attention mechanisms.
AB - We present a convolutional neural network (CNN) equipped with a novel and efficient adaptive dual attention module (ADAM) for automated skin lesion segmentation from dermoscopic images, which is an essential yet challenging step for the development of a computer-assisted skin disease diagnosis system. The proposed ADAM has three compelling characteristics. First, we integrate two global context modeling mechanisms into the ADAM, one aiming at capturing the boundary continuity of skin lesion by global average pooling while the other dealing with the shape irregularity by pixel-wise correlation. In this regard, our network, thanks to the proposed ADAM, is capable of extracting more comprehensive and discriminative features for recognizing the boundary of skin lesions. Second, the proposed ADAM supports multi-scale resolution fusion, and hence can capture multi-scale features to further improve the segmentation accuracy. Third, as we harness a spatial information weighting method in the proposed network, our method can reduce a lot of redundancies compared with traditional CNNs. The proposed network is implemented based on a dual encoder architecture, which is able to enlarge the receptive field without greatly increasing the network parameters. In addition, we assign different dilation rates to different ADAMs so that it can adaptively capture distinguishing features according to the size of a lesion. We extensively evaluate the proposed method on both ISBI2017 and ISIC2018 datasets and the experimental results demonstrate that, without using network ensemble schemes, our method is capable of achieving better segmentation performance than state-of-the-art deep learning models, particularly those equipped with attention mechanisms.
KW - adaptive dual attention module
KW - deep learning
KW - dual encoder architecture
KW - global context modeling
KW - Skin lesion segmentation
UR - http://www.scopus.com/inward/record.url?scp=85098852434&partnerID=8YFLogxK
U2 - 10.1109/TMI.2020.3027341
DO - 10.1109/TMI.2020.3027341
M3 - Journal article
C2 - 32986547
AN - SCOPUS:85098852434
SN - 0278-0062
VL - 40
SP - 357
EP - 370
JO - IEEE Transactions on Medical Imaging
JF - IEEE Transactions on Medical Imaging
IS - 1
M1 - 9207942
ER -