Scientific Reports

Table 3 Performance comparison of deep learning models with mean ± SD values for precision, sensitivity, F1 score, and accuracy, evaluated using the validation dataset for classifying CVI & DVT, lymphedema, normal, and systemic disease, based on k-fold cross-validation.

From: Deep learning-based classification of lymphedema and other lower limb edema diseases using clinical images

Architecture	Method	Lymphedema			CVI&DVT			Normal			Systemic Disease			All Class
Architecture	Method	Precision	Sensitivity	F1	Precision	Sensitivity	F1	Precision	Sensitivity	F1	Precision	Sensitivity	F1	Precision	Sensitivity	F1	Accuracy
CNN	AlexNet	0.886 ± 0.035	0.890 ± 0.030	0.888 ± 0.025	0.844 ± 0.045	0.837 ± 0.031	0.840 ± 0.026	0.805 ± 0.039	0.792 ± 0.045	0.798 ± 0.039	0.957 ± 0.034	0.972 ± 0.018	0.964 ± 0.024	0.870 ± 0.023	0.870 ± 0.023	0.866 ± 0.024	0.870 ± 0.023
	GoogLeNet	0.951 ± 0.025	0.930 ± 0.023	0.940 ± 0.022	0.927 ± 0.036	0.887 ± 0.040	0.905 ± 0.026	0.877 ± 0.044	0.887 ± 0.024	0.881 ± 0.025	0.949 ± 0.038	0.997 ± 0.006	0.972 ± 0.023	0.925 ± 0.017	0.923 ± 0.017	0.923 ± 0.017	0.923 ± 0.017
	ResNet50	0.954 ± 0.025	0.936 ± 0.025	0.945 ± 0.019	0.941 ± 0.032	0.887 ± 0.053	0.911 ± 0.022	0.881 ± 0.035	0.910 ± 0.022	0.895 ± 0.020	0.957 ± 0.017	0.992 ± 0.011	0.974 ± 0.012	0.932 ± 0.013	0.930 ± 0.013	0.930 ± 0.014	0.930 ± 0.013
	VGG16	0.938 ± 0.024	0.917 ± 0.030	0.927 ± 0.018	0.942 ± 0.018	0.895 ± 0.023	0.918 ± 0.015	0.867 ± 0.036	0.916 ± 0.029	0.890 ± 0.020	0.970 ± 0.016	0.975 ± 0.018	0.972 ± 0.017	0.927 ± 0.013	0.925 ± 0.014	0.925 ± 0.014	0.925 ± 0.014
	MobileNetV3	0.856 ± 0.076	0.836 ± 0.054	0.841 ± 0.014	0.772 ± 0.066	0.735 ± 0.085	0.747 ± 0.033	0.724 ± 0.043	0.720 ± 0.061	0.719 ± 0.027	0.916 ± 0.022	0.958 ± 0.018	0.936 ± 0.014	0.814 ± 0.017	0.809 ± 0.013	0.808 ± 0.016	0.809 ± 0.013
	DenseNet169	0.956 ± 0.019	0.945 ± 0.024	0.950 ± 0.010	0.945 ± 0.026	0.936 ± 0.022	0.940 ± 0.022	0.909 ± 0.024	0.916 ± 0.024	0.912 ± 0.017	0.976 ± 0.018	0.986 ± 0.009	0.981 ± 0.012	0.945 ± 0.009	0.945 ± 0.009	0.945 ± 0.009	0.945 ± 0.009
	SqueezeNet	0.910 ± 0.028	0.864 ± 0.052	0.885 ± 0.025	0.873 ± 0.034	0.807 ± 0.036	0.838 ± 0.022	0.780 ± 0.046	0.858 ± 0.044	0.816 ± 0.041	0.948 ± 0.020	0.958 ± 0.018	0.953 ± 0.011	0.875 ± 0.021	0.870 ± 0.022	0.871 ± 0.022	0.870 ± 0.022
	EfficientNetV2	0.964 ± 0.008	0.943 ± 0.023	0.953 ± 0.013	0.936 ± 0.023	0.920 ± 0.014	0.928 ± 0.010	0.912 ± 0.014	0.914 ± 0.021	0.913 ± 0.013	0.952 ± 0.013	0.992 ± 0.011	0.971 ± 0.009	0.941 ± 0.009	0.941 ± 0.009	0.941 ± 0.009	0.941 ± 0.009
Transformer	ViT	0.921 ± 0.038	0.877 ± 0.026	0.898 ± 0.026	0.882 ± 0.025	0.823 ± 0.046	0.851 ± 0.030	0.784 ± 0.020	0.842 ± 0.030	0.812 ± 0.018	0.929 ± 0.035	0.958 ± 0.015	0.943 ± 0.021	0.877 ± 0.014	0.874 ± 0.013	0.874 ± 0.014	0.874 ± 0.013
	TnT	0.897 ± 0.017	0.873 ± 0.039	0.884 ± 0.021	0.891 ± 0.034	0.807 ± 0.059	0.844 ± 0.023	0.789 ± 0.027	0.856 ± 0.029	0.820 ± 0.016	0.943 ± 0.021	0.961 ± 0.020	0.952 ± 0.014	0.876 ± 0.011	0.873 ± 0.010	0.873 ± 0.010	0.873 ± 0.010
	Swin	0.965 ± 0.022	0.893 ± 0.046	0.927 ± 0.029	0.914 ± 0.036	0.900 ± 0.027	0.907 ± 0.025	0.852 ± 0.025	0.912 ± 0.036	0.880 ± 0.015	0.960 ± 0.029	0.978 ± 0.017	0.969 ± 0.016	0.922 ± 0.008	0.919 ± 0.007	0.919 ± 0.007	0.919 ± 0.007
	CvT	0.804 ± 0.043	0.790 ± 0.034	0.795 ± 0.014	0.701 ± 0.053	0.569 ± 0.086	0.624 ± 0.058	0.639 ± 0.045	0.713 ± 0.073	0.670 ± 0.022	0.878 ± 0.020	0.917 ± 0.045	0.896 ± 0.022	0.752 ± 0.023	0.748 ± 0.022	0.745 ± 0.022	0.748 ± 0.022
	PiT	0.959 ± 0.016	0.906 ± 0.053	0.931 ± 0.026	0.950 ± 0.014	0.895 ± 0.045	0.921 ± 0.026	0.859 ± 0.027	0.928 ± 0.023	0.892 ± 0.016	0.957 ± 0.005	0.983 ± 0.010	0.970 ± 0.006	0.929 ± 0.010	0.927 ± 0.011	0.927 ± 0.011	0.927 ± 0.011
	CCT	0.942 ± 0.032	0.915 ± 0.038	0.927 ± 0.006	0.886 ± 0.037	0.859 ± 0.055	0.871 ± 0.021	0.851 ± 0.033	0.874 ± 0.042	0.861 ± 0.013	0.942 ± 0.030	0.964 ± 0.023	0.952 ± 0.009	0.905 ± 0.005	0.902 ± 0.004	0.902 ± 0.004	0.902 ± 0.004
	MaxViT	0.963 ± 0.018	0.941 ± 0.026	0.951 ± 0.009	0.941 ± 0.029	0.942 ± 0.031	0.941 ± 0.018	0.906 ± 0.025	0.910 ± 0.030	0.908 ± 0.021	0.973 ± 0.022	0.992 ± 0.011	0.982 ± 0.015	0.945 ± 0.011	0.944 ± 0.011	0.944 ± 0.011	0.944 ± 0.011
	DaViT	0.979 ± 0.017	0.906 ± 0.040	0.940 ± 0.025	0.947 ± 0.025	0.936 ± 0.036	0.941 ± 0.025	0.869 ± 0.023	0.941 ± 0.019	0.904 ± 0.012	0.978 ± 0.018	0.981 ± 0.021	0.979 ± 0.013	0.942 ± 0.014	0.939 ± 0.014	0.939 ± 0.014	0.939 ± 0.014

Back to article page

Search

Advanced search

Quick links