DFA-Net: Dual multi-scale feature aggregation network for vessel segmentation in X-ray digital subtraction angiography

Journal of Big Data

Table 2 Quantitative comparisons acquired from different segmentation strategies

Method	IoU	Dice	Accuracy	Specificity	Precision
U-net (2015)	0.7377	0.8491	0.9798	0.9896	0.8538
Deeplabv3+ (2018)	0.6731	0.8046	0.9709	0.9766	0.7331
U-net++ (2018)	0.7852	0.8797	0.9836	0.9899	0.8651
SA-Unet (2020)	0.7201	0.8373	0.9767	0.9829	0.7894
U-net 3+ (2020)	0.7541	0.8598	0.9800	0.9848	0.8123
CMU-Net (2022)	0.7746	0.8730	0.9822	0.9873	0.8379
CA-Net (2023)	0.7881	0.8815	0.9837	0.9896	0.8623
DFA-Net (ours)	0.8011	0.8896	0.9852	0.9924	0.8941