Skip to main content

Table 3 Quantitative comparison of proposed algorithm with existing state-of-the-art machine learning-based image defogging

From: Segmentation-based image defogging using modified dark channel prior

Quality MetricsDehaze-Net [37]MSCNN [36]AoD-Net [38]Proposed IProposed II
Image 1
SSIM0.73620.80400.82450.862330.8916
PSNR19.058317.40120.925219.401420.9254
NIQE2.62462.4542.29923.10893.0930
BRISQUE20.274127.401218.929518.455618.5600
Image 2
SSIM0.8630.86450.94320.934170.9032
PSNR19.576919.785321.758222.47720.5093
NIQE3.99584.055520.63514.03734.3324
BRISQUE22.39124.961418.527417.747431.0225
Image 3
SSIM0.88590.86350.86230.916040.9184
PSNR24.449322.131025.509322.965525.0210
NIQE3.53453.26234.14463.25264.3324
BRISQUE17.485528.688518.280916.870522.5321