Feature Selection Based on Dialectical Optimization Algorithm for Breast Lesion Classification in Thermographic Images

Pereira, Jessiane Mônica Silva; Santana, Maíra Araújo de; Lima, Clarisse Lins de; Lima, Ricardo Carvalho; Lima, Sidney Marlon Lopes de; Santos, Wellington Pinheiro dos

doi:10.4018/978-1-7998-3456-4.ch004

Cited by 4 publications

(3 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Through the error obtained, it is possible to adjust the weights and iteratively minimize the 68 error. MLPs have been commonly used in medical applications, such as in the diagnosis of cancer [69][70][71][72][73][74][75] , diabetes 76 , multiple sclerosis 77 , and in the monitoring and diagnosis of Covid-19 [78][79][80] .…”

Section: Sparse Representation Classificationmentioning

confidence: 99%

Motor imagery classification using sparse representations: an exploratory study

de Menezes,

Gomes,

de Carvalho Hazin

et al. 2023

Sci Rep

View full text Add to dashboard Cite

The non-stationary nature of the EEG signal poses challenges for the classification of motor imagery. sparse representation classification (SRC) appears as an alternative for classification of untrained conditions and, therefore, useful in motor imagery. Empirical mode decomposition (EMD) deals with signals of this nature and appears at the rear of the classification, supporting the generation of features. In this work we evaluate the combination of these methods in a multiclass classification problem, comparing them with a conventional method in order to determine if their performance is regular. For comparison with SRC we use multilayer perceptron (MLP). We also evaluated a hybrid approach for classification of sparse representations with MLP (RSMLP). For comparison with EMD we used filtering by frequency bands. Feature selection methods were used to select the most significant ones, specifically Random Forest and Particle Swarm Optimization. Finally, we used data augmentation to get a more voluminous base. Regarding the first dataset, we observed that the classifiers that use sparse representation have results equivalent to each other, but they outperform the conventional MLP model. SRC and SRMLP achieve an average accuracy of $$75.95\%$$ 75.95 % and $$82.51\%$$ 82.51 % respectively while the MLP is $$72.38\%$$ 72.38 % , representing a gain between $$4.93\%$$ 4.93 % and $$14\%$$ 14 % . The use of EMD in relation to other feature processing techniques is not superior. However, EMD does not influence negatively, there is an opportunity for improvement. Finally, the use of data augmentation proved to be important to obtain relevant results. In the second dataset, we did not observe the same results. Models based on sparse representation (SRC, SRMLP, etc.) have on average a performance close to other conventional models, but without surpassing them. The best sparse models achieve an average accuracy of $$95.43\%$$ 95.43 % among the subjects in the base, while other model reach $$98.33\%$$ 98.33 % . The improvement of self-adaptive mechanisms that respond efficiently to the user’s context is a good way to achieve improvements in motor imagery applications. However, other scenarios should be investigated, since the advantage of these methods was not proven in all datasets studied. There is still room for improvement, such as optimizing the dictionary of sparse representation in the context of motor imagery. Investing efforts in synthetically increasing the training base has also proved important to reduce the costs of this group of applications.

show abstract

Section: Sparse Representation Classificationmentioning

confidence: 99%

Motor imagery classification using sparse representations: an exploratory study

de Menezes,

Gomes,

de Carvalho Hazin

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Existen varias técnicas de pre procesamiento que se han utilizado para eliminar ruido y mejorar la imagen como: segmentación basada en bordes, segmentación orientada a regiones, entre otros que permiten eliminar los factores de interferencia de las imá-genes de dermatoscopia. Pereira et al [6], describe que, en la detección del cáncer de piel se puede determinar tres puntos clave: mejora de la imagen, restauración de la imagen y eliminación de partes sin relevancia, para lograr un correcto procesamiento de la imagen.…”

Section: Proposal Of a Machine Learning Application With Artificial V...unclassified

Propuesta de una aplicación de aprendizaje automático con visión artificial como herramienta de apoyo para la detección de melanomas benignos y malignos.

Quintanilla,

Tamayo,

Suntaxi

et al. 2022

GEEKS

View full text Add to dashboard Cite

El cáncer de piel es una de las enfermedades que más muertes causa mundialmente, siendo especialmente conocida por la agresividad con la que esta enfermedad se propaga. Sin embargo, la tasa de supervivencia es alta si los melanomas son detectados en etapas iniciales, siendo diagnosticados inicialmente por su tamaño, forma y coloración antes de continuar con procesos invasivos. En el presente paper se muestra una propuesta de aplicación como una herramienta aplicando técnicas de machine learning y procesamiento de imágenes que sea capaz de facilitar el diagnóstico de los melanomas categorizándolos por las características presentadas por la lesión analizada, con el propósito de ayudar al especialista en el diagnóstico del tratamiento que el paciente con esta lesión cutánea deba tomar.

show abstract

“…An improved method of classifying breast cancer using deep learning was recently presented by the authors of [24]. The authors of [25] proposed dialectical feature selection to improve breast cancer classification; however, these methods run into the issue of stopping after the ideal values have been retrieved.…”

Section: Introductionmentioning

confidence: 99%

Classification of Breast Cancer Using Transfer Learning and Advanced Al-Biruni Earth Radius Optimization

et al. 2023

View full text Add to dashboard Cite

Breast cancer is one of the most common cancers in women, with an estimated 287,850 new cases identified in 2022. There were 43,250 female deaths attributed to this malignancy. The high death rate associated with this type of cancer can be reduced with early detection. Nonetheless, a skilled professional is always necessary to manually diagnose this malignancy from mammography images. Many researchers have proposed several approaches based on artificial intelligence. However, they still face several obstacles, such as overlapping cancerous and noncancerous regions, extracting irrelevant features, and inadequate training models. In this paper, we developed a novel computationally automated biological mechanism for categorizing breast cancer. Using a new optimization approach based on the Advanced Al-Biruni Earth Radius (ABER) optimization algorithm, a boosting to the classification of breast cancer cases is realized. The stages of the proposed framework include data augmentation, feature extraction using AlexNet based on transfer learning, and optimized classification using a convolutional neural network (CNN). Using transfer learning and optimized CNN for classification improved the accuracy when the results are compared to recent approaches. Two publicly available datasets are utilized to evaluate the proposed framework, and the average classification accuracy is 97.95%. To ensure the statistical significance and difference between the proposed methodology, additional tests are conducted, such as analysis of variance (ANOVA) and Wilcoxon, in addition to evaluating various statistical analysis metrics. The results of these tests emphasized the effectiveness and statistical difference of the proposed methodology compared to current methods.

show abstract

Feature Selection Based on Dialectical Optimization Algorithm for Breast Lesion Classification in Thermographic Images

Cited by 4 publications

References 22 publications

Motor imagery classification using sparse representations: an exploratory study

Motor imagery classification using sparse representations: an exploratory study

Propuesta de una aplicación de aprendizaje automático con visión artificial como herramienta de apoyo para la detección de melanomas benignos y malignos.

Classification of Breast Cancer Using Transfer Learning and Advanced Al-Biruni Earth Radius Optimization

Contact Info

Product

Resources

About