Melanoma Detection by Means of Multiple Instance Learning

Astorino, Annabella; Fuduli, Antonio; Veltri, Pierangelo; Vocaturo, Eugenio

doi:10.1007/s12539-019-00341-y

Cited by 44 publications

(21 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…All the images of the database were selected on the basis of their quality, resolution and dermoscopic features. Good results were obtained also in [7] on the same data set Ph 2 , On the other hand, the objective in [34] was to select the most important image features usable in melanoma detection. In this case, differently from [33] where a high quality data set was adopted, the authors performed their experimentations on a data set constituted by plain photographs publicly available from two online databases https://www.dermquest.coml and http://www.dermins.netl, obtaining very good results in terms of sensitivity (96.77%) and, consequently, in terms of F-score (90.91%).…”

Section: Numerical Experimentationsmentioning

confidence: 91%

“…In this section a preliminary numerical comparison between two different approaches (Support Vector Machine and Multiple Instance Learning) is reported, using color and color/texture features. This study starts from some considerations related to the works [7,33,34]. In [33] the authors analyzed the role played by the color and the texture features, showing empirically that using only the color features outperforms the use of the texture features: very good results were obtained by means of different type of classifiers on a image data set drawn from the Ph 2 database [8], containing 200 melanocytic lesions images (80 common nevi, 80 atypical nevi and 40 melanomas).…”

Section: Numerical Experimentationsmentioning

confidence: 99%

“…The MIL paradigm applies very well to image classification since in general an image (bag) is classified on the basis of some portions (instances) of it. About the MIL-RL code, the same Matlab implementation adopted in [7] were used, while, for the SVM approaches with both linear and RBF kernels, the fitcsvm and predict subroutines for training the classifier and for computing the testing correctness, respectively, were utilized. Such subroutines are included in the Matlab optimization package.…”

Section: Numerical Experimentationsmentioning

confidence: 99%

“…Here a preliminary comparison between two different approaches (Support Vector Machine and Multiple Instance Learning) is proposed, pointing out the key role played by the feature selection (color and texture). In particular, this study was inspired by the good results obtained in [7], where the authors applied Multiple Instance Learning techniques for classifying some medical dermoscopic images drawn from the Ph 2 database [8], by using only color features. Our numerical experimentations, reported in this paper, showed that using only color features is not at all high-performance when the classification process is performed on a data base constituted by common plain photographies.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Melanoma detection using color and texture features in computer vision systems

Fuduli¹,

Veltri²,

Vocaturo³

et al. 2019

Adv. sci. technol. eng. syst. j.

Self Cite

View full text Add to dashboard Cite

All forms of skin cancer are becoming widespread. These forms of cancer, and melanoma in particular, are insidious and aggressive and if not treated promptly can be lethal to humans. Effective treatment of skin lesions depends strongly on the timeliness of the diagnosis: for this reason, artificial vision systems are required to play a crucial role in supporting the diagnosis of skin lesions. This work offers insights into the state of the art in the field of melanoma image classification. We include a numerical section where a preliminary analysis of some classification techniques is performed, using color and texture features on a data set constituted by plain photographies, to which no pre-processing technique has been applied. This is motivated by the necessity to open new horizons in creating self-diagnosis systems for accessible skin lesions, due also to a huge innovation of cameras, smartphones technology and wearable devices.

show abstract

Section: Numerical Experimentationsmentioning

confidence: 91%

Section: Numerical Experimentationsmentioning

confidence: 99%

Section: Numerical Experimentationsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Melanoma detection using color and texture features in computer vision systems

Fuduli¹,

Veltri²,

Vocaturo³

et al. 2019

Adv. sci. technol. eng. syst. j.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Some MIL applications are image classification [5][6][7][8], drug discovery [9,10], classification of text documents [11], bankruptcy prediction [12], and speaker identification [13].…”

Section: Introductionmentioning

confidence: 99%

SVM-Based Multiple Instance Classification via DC Optimization

et al. 2019

Self Cite

View full text Add to dashboard Cite

A multiple instance learning problem consists of categorizing objects, each represented as a set (bag) of points. Unlike the supervised classification paradigm, where each point of the training set is labeled, the labels are only associated with bags, while the labels of the points inside the bags are unknown. We focus on the binary classification case, where the objective is to discriminate between positive and negative bags using a separating surface. Adopting a support vector machine setting at the training level, the problem of minimizing the classification-error function can be formulated as a nonconvex nonsmooth unconstrained program. We propose a difference-of-convex (DC) decomposition of the nonconvex function, which we face using an appropriate nonsmooth DC algorithm. Some of the numerical results on benchmark data sets are reported.

show abstract

An all‐inclusive computer‐aided melanoma diagnosis based on soft computing

Jiao

Zhao

et al. 2022

Int J Imaging Syst Tech

View full text Add to dashboard Cite

This paper provides a different optimized approach for melanoma diagnosis from the inputted dermoscopy images. The technique is a pipeline technique with four main steps, including noise reduction, lesion segmentation, feature selection, and final classification. For decreasing the complexity of the feature extraction stage, Fuzzy C-means has been used. The classifier has been improved based on a developed decision tree. The modification of the classifier is based on a new enhanced design of a metaheuristic, called Quantum Fluid Search Optimizer. The efficiency of the suggested technique is calculated by considering some measurement indicators and their achievements are compared with five other latest methods. The results showed the maximum accuracy equal to 94.12% with the highest precision being achieved by the proposed method. The results also indicate that the proposed method with the highest value of 91.18% sensitivity against the other techniques, provides the highest reliability.

show abstract

Melanoma Detection by Means of Multiple Instance Learning

Cited by 44 publications

References 27 publications

Melanoma detection using color and texture features in computer vision systems

Melanoma detection using color and texture features in computer vision systems

SVM-Based Multiple Instance Classification via DC Optimization

An all‐inclusive computer‐aided melanoma diagnosis based on soft computing

Contact Info

Product

Resources

About