Source-linked AI summary
Image Classification of Melanoma, Nevus and Seborrheic Keratosis by Deep Neural Network Ensemble
Kazuhisa Matsunaga, Akira Hamada, Akane Minagawa, Hiroshi Koga
TL;DR
The paper addresses lesion classification in Part 3 of the ISIC-ISBI 2017 challenge. It presents a classification system and machine-learning approach, reporting performance that significantly outperformed the previous state of the art despite no lesion segmentation or cropping.
Problem
The paper addresses the defined lesion-classification task in Part 3 of the ISIC-ISBI 2017 challenge.
Source: §1 · p. 1Method
The authors present a proposed classification system and machine-learning approach, together with evaluation results.
Source: §1 · p. 1Results
Takeaways & Limitations
The reported results indicate that the proposed method achieved strong challenge performance without using lesion segmentation or cropping.
Source: §5 · p. 4The method did not use lesion segmentation or cropping, and the authors expect reliable segmentation to improve results to a certain extent.
Source: §5 · p. 4
Abstract
from arXiv · showhide
This short paper reports the method and the evaluation results of Casio and Shinshu University joint team for the ISBI Challenge 2017 - Skin Lesion Analysis Towards Melanoma Detection - Part 3: Lesion Classification hosted by ISIC. Our online validation score was 0.958 with melanoma classifier AUC 0.924 and seborrheic keratosis classifier AUC 0.993.
1. Introduction
The challenge classified skin-lesion images into melanoma, nevus, and seborrheic keratosis using two binary classifiers, with participants ranked by mean AUC. The report presents the team's classification system and evaluation results.
- The task classified images into melanoma, nevus, and seborrheic keratosis classes.Source: §1 · p. 1
- Two binary classifiers addressed melanoma versus the rest and seborrheic keratosis versus the rest.Source: §1 · p. 1
- Participants were ranked by the mean of the two classifier area under the ROC curves.Source: §1 · p. 1
- The task allowed external training data and age/sex information tagged with some provided samples.Source: §1 · p. 1
- The report presents the proposed classification system, machine learning approach, and evaluation results.Source: §1 · p. 1
2. Proposed Classification System
The proposed system normalizes lesion images, processes transformed inputs with CNN ensembles for two binary tasks, and optionally uses age/sex information for SK classification. It also integrates SK outputs into MM classification because the SK classifier was considered more reliable.
- Input images undergo luminance, color-balance, and color-constancy normalization before classification.Source: §2 · p. 1
- Normalized images feed separate SK-versus-rest and MM-versus-rest base classifiers with identical composition.Source: §2 · p. 1
- Geometric image transformations are processed in parallel by CNN ensembles that output prediction values in [0.0,1.0].Source: §2 · p. 1
- The system uses 50-layer ResNet models implemented in Keras with modifications by the authors.Source: §2 · p. 1
- Age/sex information may be thresholded before output for SK classification, while cross-validation found no significant increase for MM classification.Source: §2 · p. 1
- The authors integrated SK-classifier information into MM classification because the SK classifier was more reliable despite training on the same sample set.Source: §2 · pp. 1–2
3. Machine Learning
The machine-learning pipeline augments provided data with external ISIC Archive samples, fine-tunes pretrained CNNs, and selects CNN combinations through cross-validation.
- The training data comprised 374 MM, 254 SK, and 1372 NCN provided samples.Source: §3 · p. 3
- External training data added 409 MM, 66 SK, and 969 NCN samples from an ISIC Archive subset.Source: §3 · p. 3
- CNNs were fine-tuned from an initial Keras model pretrained for generic object recognition.Source: §3 · p. 3
- Different optimization methods were tested, and the best CNN combination was selected through cross-validation.Source: §3 · p. 3
- The optimization methods were RMSProp and AdaGrad.Source: §3 · p. 3
4. Results
The proposed method was evaluated on ISIC-ISBI Challenge 2017 validation data and Challenge 2016 test data. External data and complementary classifier use improved selected AUC results, while age/sex information helped SK classification but not MM classification.
- Age/sex information improved SK AUC from 0.957 to 0.960 in cross-validation but produced no improvement for MM.Source: §4 · p. 3
- The complementary use of the SK classifier improved MM AUC from 0.917 to 0.924 on the validation set.Source: §4 · p. 3
- Challenge 2016 Part 3 evaluated malignant-versus-benign binary classification using the 2016 training set with external samples.Source: §4 · p. 3
- The reported evaluation included Challenge 2017 validation-set scores and Challenge 2016 test-set scores, with final results presented in Table 1.Source: §4 · p. 3, Table 1 · p. 3
5. Conclusions and Future Work
The proposed method significantly outperformed last year’s state-of-the-art despite using no lesion segmentation or cropping. The authors identify segmentation, age/sex handling, and multi-class integration as areas requiring further consideration.
- The proposed method significantly outperformed last year’s state-of-the-art.Source: §5 · p. 4
- The system did not utilize lesion segmentation or cropping, although reliable segmentation could improve results to a certain extent.Source: §5 · p. 4
- Age/sex information had a slight effect on seborrheic keratosis classification.Source: §5 · p. 4
- Practical use may require more careful implementation than thresholding because seborrheic keratosis is rare at young ages but still possible.Source: §5 · p. 4
- The multi-class integration may require greater mathematical sophistication and consideration of how to treat complex diseases.Source: §5 · p. 4