Final published version
Licence: CC BY-NC-ND: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - Deep learning in diabetic foot ulcers detection
T2 - A comprehensive evaluation
AU - Yap, Moi Hoon
AU - Hachiuma, R
AU - Alavi, A
AU - Brüngel, R
AU - Cassidy, Bill
AU - Goyal, M
AU - Zhu, H
AU - Rückert, J
AU - Olshansky, M
AU - Huang, X
AU - Saito, H
AU - Hassanpour, S
AU - Friedrich, C M
AU - Ascher, D B
AU - Song, A
AU - Kajita, H
AU - Gillespie, David
AU - Reeves, Neil
AU - Pappachan, J M
AU - O'Shea, C
AU - Frank, E
PY - 2021/8/31
Y1 - 2021/8/31
N2 - There has been a substantial amount of research involving computer methods and technology for the detection and recognition of diabetic foot ulcers (DFUs), but there is a lack of systematic comparisons of state-of-the-art deep learning object detection frameworks applied to this problem. DFUC2020 provided participants with a comprehensive dataset consisting of 2,000 images for training and 2,000 images for testing. This paper summarizes the results of DFUC2020 by comparing the deep learning-based algorithms proposed by the winning teams: Faster R–CNN, three variants of Faster R–CNN and an ensemble method; YOLOv3; YOLOv5; EfficientDet; and a new Cascade Attention Network. For each deep learning method, we provide a detailed description of model architecture, parameter settings for training and additional stages including pre-processing, data augmentation and post-processing. We provide a comprehensive evaluation for each method. All the methods required a data augmentation stage to increase the number of images available for training and a post-processing stage to remove false positives. The best performance was obtained from Deformable Convolution, a variant of Faster R–CNN, with a mean average precision (mAP) of 0.6940 and an F1-Score of 0.7434. Finally, we demonstrate that the ensemble method based on different deep learning methods can enhance the F1-Score but not the mAP.
AB - There has been a substantial amount of research involving computer methods and technology for the detection and recognition of diabetic foot ulcers (DFUs), but there is a lack of systematic comparisons of state-of-the-art deep learning object detection frameworks applied to this problem. DFUC2020 provided participants with a comprehensive dataset consisting of 2,000 images for training and 2,000 images for testing. This paper summarizes the results of DFUC2020 by comparing the deep learning-based algorithms proposed by the winning teams: Faster R–CNN, three variants of Faster R–CNN and an ensemble method; YOLOv3; YOLOv5; EfficientDet; and a new Cascade Attention Network. For each deep learning method, we provide a detailed description of model architecture, parameter settings for training and additional stages including pre-processing, data augmentation and post-processing. We provide a comprehensive evaluation for each method. All the methods required a data augmentation stage to increase the number of images available for training and a post-processing stage to remove false positives. The best performance was obtained from Deformable Convolution, a variant of Faster R–CNN, with a mean average precision (mAP) of 0.6940 and an F1-Score of 0.7434. Finally, we demonstrate that the ensemble method based on different deep learning methods can enhance the F1-Score but not the mAP.
U2 - 10.1016/j.compbiomed.2021.104596
DO - 10.1016/j.compbiomed.2021.104596
M3 - Journal article
C2 - 34247133
VL - 135
JO - Computers in biology and medicine
JF - Computers in biology and medicine
SN - 0010-4825
M1 - 104596
ER -