Skip to main content

Table 1 Various network model versus dermatologist’s precision, sensitivity, and specificity values

From: Automatic identification of benign pigmented skin lesions from clinical images using deep convolutional neural network

Diseases

SSD

Faster R-CNN

YOLOv4

YOLOv5

DOCTOR

Precision

Sensitivity

Specificity

Precision

Sensitivity

Specificity

Precision

Sensitivity

Specificity

Precision

Sensitivity

Specificity

Precision

Sensitivity

Specificity

Sl

0.938

0.882

0.940

0.955

0.840

0.961

0.920

0.92

0.922

0.920

0.92

0.922

1.000

0.967

1.00

Fre

0.960

0.923

0.960

0.909

0.980

0.902

0.959

0.904

0.96

1.000

0.942

1.00

0.994

0.994

0.980

Mel

0.850

0.927

0.836

0.877

0.862

0.865

0.877

0.877

0.870

0.930

0.93

0.925

0.978

0.989

0.926

Caf

0.882

0.94

0.887

0.980

0.990

0.941

0.927

1.00

0.922

0.944

1.00

0.922

1.000

0.993

1.00

Ota

0.979

1.00

0.980

0.978

0.938

0.980

0.960

1.00

0.961

0.980

0.98

0.980

1.000

1.00

1.00

Hori

0.906

0.980

0.906

0.940

0.940

0.942

0.958

0.958

0.961

0.962

1.00

0.962

0.993

0.981

0.980

Average

0.919

0.942

0.918

0.940

0.925

0.932

0.934

0.943

0.933

0.956

0.962

0.952

0.994

0.987

0.981