Cite

Figure 1

Skin cancer stages (source: teleskin.org)
Skin cancer stages (source: teleskin.org)

Figure 2

The exploded view of the probe containing electrical impedance electrodes as well as temperature sensors.
The exploded view of the probe containing electrical impedance electrodes as well as temperature sensors.

Figure 3

Design of the probe in SolidWorks and the prototype
Design of the probe in SolidWorks and the prototype

Figure 4

Use of probe and the pressure sensors over human tissue
Use of probe and the pressure sensors over human tissue

Figure 5

The Debye case of the Cole model circuit (α=1).
The Debye case of the Cole model circuit (α=1).

Figure 6

Imaginary part of admittance versus its real part (Cole plot).
Imaginary part of admittance versus its real part (Cole plot).

Figure 7

Paths of high and low frequency currents in a biological tissue.
Paths of high and low frequency currents in a biological tissue.

Figure 8

Admittance plot of the tumorous subjects and the contralateral healthy part.
Admittance plot of the tumorous subjects and the contralateral healthy part.

Figure 9

Comparison of the performance of 3 classification methods on the data.
Comparison of the performance of 3 classification methods on the data.

Case studies recruited for testing the proposed method on cancerous subjects and their information

Case Study No.AgeSexTumor TypeTumor Position
Subject 190FemaleBCC (Basal Cell Carcinoma)Left side of nose along sidewall
Subject 297FemaleSCC (Squamous Cell Carcinoma)Left cheek center
Subject 381MaleSCC (Squamous Cell Carcinoma)Right Temple of head
Subject 493FemaleSCC (Squamous Cell Carcinoma)Left cheek
Subject 587FemaleBCC (Basal Cell Carcinoma)Left cheek under the eye
Subject 693FemaleBCC (Basal Cell Carcinoma)Left mid neck
Subject 787MaleBCC (Basal Cell Carcinoma)Right cheek
Subject 892MaleSCC (Squamous Cell Carcinoma)Forearm
Subject 966FemaleBCC (Basal Cell Carcinoma)Upper left cheek
Subject 1065MaleBCC (Basal Cell Carcinoma)Left nasal wing

The performance of various classification methods on raw and processed data.

Sensitivity %Specificity %Accuracy %Resubstitution %Cross Validation Error %
Raw DataNB Gaussian7.145031.2568.7554.17
NB Kernel12.512.512.587.587.5
Extracted FeaturesNB Gaussian38.4673.6859.3840.6331.2
NB Kernel44.4473.9165.6334.3833.3
SVM2058.8240.6359.3853.12
Normalized Extracted FeaturesNB Gaussian66.679081.2518.7516.67
NB Kernel7086.3681.2518.7513.12
SVM508068.7531.2518.75