Segmentation of Melanocytic Lesion Images Using Gamma Correction with Clustering of Keypoint Descriptors

. 2021 Jul 29;11(8):1366. doi: 10.3390/diagnostics11081366

Algorithm 1. Preprocessing.

I_{a}

= smoothening of image

I_{s}

using Ambrosio-Tortorelli [58] minimizer

I_{g}

= gray level of

I_{a}

L_{f}

= fast line detections using length threshold of 20
Len

(L_{f})

= size of detected L_{f}

if Len

(L_{f}) \geq 1

create a 2-d kernel

k_{b m}

based on the Len

(L_{f})

k_{b m} = 3 \times 3

If 20 ≥ Len

(L_{f})

≥ 1

k_{b m} = 7 \times 7

35 ≥ Len

(L_{f})

k_{b m} = 11 \times 11

If 50 ≥ Len

(L_{f})

k_{b m} = 15 \times 15

If Len

(L_{f})

I_{b m}

= morphology blackhat of

I_{g}

using

k_{b m}

I_{b t}

= morphology binary threshold of

I_{b m}

I_{p}

= fast marching inpaint [61] of

I_{b t}

else

I_{p}

= initialize as

I_{a}