SLC_ShiftingEnsemble

This repository contains the source code of the paper Skin lesion classification by ensembles of deep convolutional networks and regularly spaced shifting.

This code executes the Shifted GoogLeNet+MobileNetV2 method for the HAM10000 dataset. The contents of this code are provided without any warranty. They are intended for evaluational purposes only.

Pre-requisites

• Matlab (tested on v2020b or earlier). Deep learning toolbox is required to load GoogLeNet and MobileNetV2.

---

Training

1. Open trainNets.m and set up the paths of the dataset
2. Run the script

---

Testing

1. Open testNetGrids.m and set up the paths of the dataset
2. Run the script

Evaluation

• computeStatsCV.m: computes the statistics of the 10-fols CV
• plotConfusionCV.m: computes the confusion matrices of the tested models
• plotModelsComparisonCV.m: plots the graph bar comparing all models

---

Citation

Please, cite this work as:

K. Thurnhofer-Hemsi, E. López-Rubio, E. Domínguez and D. A. Elizondo, "Skin Lesion Classification by Ensembles of Deep Convolutional Networks and Regularly Spaced Shifting", in IEEE Access, vol. 9, pp. 112193-112205, 2021, doi: 10.1109/ACCESS.2021.3103410. (https://ieeexplore.ieee.org/abstract/document/9508981)