A Cloud Approach for Melanoma Detection Based on Deep Learning Networks

Biasi, Luigi Di; Citarella, Alessia Auriemma; Risi, Michele; Tortora, Genoveffa

doi:10.1109/jbhi.2021.3113609

Cited by 31 publications

(24 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this previous work, we addressed the issue of Transfer Learning (TL) and the development of a more adaptable system design that can accommodate changes in training datasets. Our findings suggest that AlexNet is the most robust network in terms of TL, without data augmentation, with mean accuracies of 78% and 89% with and without Otsu segmentation, respectively [41].…”

Section: Related Workmentioning

confidence: 75%

“…In [41], we reported results that strongly suggested that the training and validation steps could suffer from intra-class dissimilarities and extra-class similarities. In particular, we rely on the hypothesis that the CNN performances can vary, even if the training, validation, and test sets vary minimally.…”

Section: Image Segmentation Methodsmentioning

confidence: 99%

“…In order to avoid biased results, we followed a similar training protocol used in [41] to consider the mean performance instead of absolute performance to make our analysis more robust. The experimental environment used was MatLab 2021b [2] .…”

Section: Image Segmentation Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Refactoring and performance analysis of the main CNN architectures: using false negative rate minimization to solve the Clinical Images Melanoma Detection Problem

Biasi

Marco

Citarella

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: Melanoma is one of the deadliest tumors in the world. Early detection is critical for first-line therapy in this tumor pathology and it remains challenging due to the need for histological analysis to ensure correctness in diagnosis. Therefore, multiple computer-aided diagnosis (CAD) systems working on melanoma images were proposed to mitigate the need of a biopsy. However, although the high global accuracy is declared in literature results, the CAD systems for the health fields must focus on the lowest false negative rate (FNR) possible to qualify as a diagnosis support system. The final goal must be to avoid classification type 2 errors to prevent life-threatening situations. Another goal could be to create an easy-to-use system for both physicians and patients. Results: To achieve the minimization of type 2 error, we performed a wide exploratory analysis of the principal convolutional neural network (CNN) architectures published for the multiple image classification problem; we adapted these networks to the melanoma clinical image binary classification problem (MCIBCP). We collected and analyzed performance data to identify the best CNN architecture, in terms of FNR, usable for solving the MCIBCP problem. Then, to provide a starting point for an easy-to-use CAD system, we used a clinical image dataset (MED-NODE) because clinical images are easier to access: they can be taken by a smartphone or other hand-size devices. Despite the lower resolution than dermoscopic images, the results in the literature would suggest that it would be possible to achieve high classification performance by using clinical images. In this work, we used MED-NODE, which consists of 170 clinical images (70 images of melanoma and 100 images of naevi). We optimized the following CNNs for the MCIBCP problem: Alexnet, DenseNet, GoogleNet Inception V3, GoogleNet, MobileNet, ShuffleNet, SqueezeNet, and VGG16. Conclusions: The results suggest that a CNN built on the VGG or AlexNet structure can ensure the lowest FNR (0.07) and (0.13), respectively. In both cases, discrete global performance is ensured. 73% (accuracy), 82% (sensitivity) and 59% (specificity) for VGG; 89% (accuracy), 87% (sensitivity) and 90% (specificity) for AlexNet.

show abstract

Section: Related Workmentioning

confidence: 75%

Section: Image Segmentation Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Refactoring and performance analysis of the main CNN architectures: using false negative rate minimization to solve the Clinical Images Melanoma Detection Problem

Biasi

Marco

Citarella

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Distributed approach deals with reducing the running time. Normal skin and cancerous skin [21] with artifacts can mislead the model. The first procedure, morphological operations, cancerous region detection.…”

Section: Literature Workmentioning

confidence: 99%

A combined approach of VGG 16 and LSTM transfer learning technique for skin melanoma classification

Arjun

Kumar

2022

ijhs

View full text Add to dashboard Cite

To distinguish melanoma from other skin illnesses, doctors examine pigmented lesions on the skin. Damaged DNA causes cells to expand uncontrollably, and the rate of growth is currently increasing rapidly. Melanoma is a kind of skin cancer induced by UV radiation from the sun, with a survival rate of about 15-20 percent. Increased UV light on the earth's surface is also aiding the spread of skin cancer throughout the globe. Melanoma is diagnosed late, resulting in severe malignancy and spread to other bodily organs such as the liver, lungs, and brain. Melanoma diagnosis from dermoscopic skin samples automatically is a difficult problem. The purpose of Computer Vision, Machine Learning, and Deep Learning in the era of digital pictures is to extract information from them and develop new knowledge. Deep convolutional neural network (DCNN) models have been extensively studied for skin disease detection, with some achieving diagnostic results that are equivalent to or even better than dermatologists. Pre-processing involves first applying a filter or kernel to reduce noise and artefacts, then trained on a variety of tiny, unbalanced datasets to ensure that the moderately complicated models outperform the bigger models. Finally, to minimise overfitting, regularization DropOut is introduced.

show abstract

“…In particular, we took into consideration how simple modification in the dataset determines a change in the accuracy of the classifiers [24]. We followed the same approach used to evaluate the performance degradation of multiple Neural Network classifiers [25] on the melanoma detection problem, allowing the training set and the test set to slightly change by randomly rebuilding the datasets for each training iteration.…”

Section: Classifiers Robustnessmentioning

confidence: 99%

ENTAIL: Yet Another Amyloid Fibrils Classifier

Citarella

Biasi

Marco

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Background: This research aims to increase our knowledge of amyloidoses. These disorders cause incorrect protein folding, affecting protein functionality (on structure). Fibrillar deposits are the basis of some wellknown diseases, such as Alzheimer, Creutzfeldt -Jakob diseases and type II diabetes. For many of these amyloid proteins, the relative precursors are known. Discovering new protein precursors involved in forming amyloid fibril deposits would improve understanding the pathological processes of amyloidoses. Results: A new classifier, called ENTAIL, was developed using over than 4000 molecular descriptors. ENTAIL was based on the Naive Bayes Classifier with Unbounded Support and Gaussian Kernel Type, with an accuracy on the test set of 81.80%, SN of 100%, SP of 63.63% and an MCC of 0.683 on a balanced dataset. Conclusions: The analysis carried out has demonstrated how, despite the various configurations of the tests, performances are superior in terms of performance on a balanced dataset.

show abstract

A Cloud Approach for Melanoma Detection Based on Deep Learning Networks

Cited by 31 publications

References 35 publications

Refactoring and performance analysis of the main CNN architectures: using false negative rate minimization to solve the Clinical Images Melanoma Detection Problem

Refactoring and performance analysis of the main CNN architectures: using false negative rate minimization to solve the Clinical Images Melanoma Detection Problem

A combined approach of VGG 16 and LSTM transfer learning technique for skin melanoma classification

ENTAIL: Yet Another Amyloid Fibrils Classifier

Contact Info

Product

Resources

About