Efficient Hyperparameter Optimization of Convolutional Neural Networks on Classification of Early Pulmonary Nodules

Lima, Lucas; Ferreira, José Raniery; Oliveira, Marcelo Costa

doi:10.1109/cbms.2019.00039

Cited by 6 publications

(7 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Among this group of studies, Choi et al 11 by applying 2, 5, and 10-fold data division protocol on SVM achieved the highest classification accuracy of 84.6%. The other group of studies [13][14][15][16][45][46][47][48][49] had explored many different well-known classification techniques, such as CNN, [13][14][15][16]46 Decision Tree, 45 Rule-Based, 45 Naïve Bayes, 45 Deep Belief Network, 46 Binary Decision Tree, 47 Random Forest, 48 and K-Nearest Neighbors 48,49 for early detection and classification of lung cancer. Among this group of studies, Lima et al 15 by exploiting the hyperparameter optimization on CNN achieved the highest classification accuracy of 88%.…”

Section: Benchmarkingmentioning

confidence: 99%

“…The other group of studies [13][14][15][16][45][46][47][48][49] had explored many different well-known classification techniques, such as CNN, [13][14][15][16]46 Decision Tree, 45 Rule-Based, 45 Naïve Bayes, 45 Deep Belief Network, 46 Binary Decision Tree, 47 Random Forest, 48 and K-Nearest Neighbors 48,49 for early detection and classification of lung cancer. Among this group of studies, Lima et al 15 by exploiting the hyperparameter optimization on CNN achieved the highest classification accuracy of 88%. Among all other well-known classification techniques, [13][14][15][16][45][46][47][48][49] CNN obtained the best classification results.…”

Section: Benchmarkingmentioning

confidence: 99%

“…The other group [13][14][15][16] uses one of the deep learning (DL) architecture, CNN, for designing a risk stratification tool that can accurately diagnose lung cancer. In 2016, Shen et al 13 designed a multi-crop CNN which can automatically extract the salient information of the nodules.…”

Section: Introductionmentioning

confidence: 99%

“…Using the LIDC-IDRI database (4581 images), CNN outperformed other networks and achieved the highest accuracy, sensitivity, and specificity of 84.15%, 83.96%, and 84.32%, respectively. In 2019, Lima et al 15 exploit hyperparameter optimization on CNN for the classification of small pulmonary nodules. By using LIDC-IDRI (1171 nodules) the study achieved the accuracy, sensitivity, and specificity of 88%, 94%, and 83% with an error rate of 12%.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Improved pulmonary lung nodules risk stratification in computed tomography images by fusing shape and texture features in a machine‐learning paradigm

Sahu

Londhe

Verma

et al. 2020

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Lung cancer is one of the most deadly cancer in both men and women. Accurate and early diagnosis of pulmonary lung nodules is critical. This study presents an accurate computer-aided diagnosis (CADx) system for risk stratification of pulmonary nodules in computed tomography (CT) lung images by fusing shape and texture-based features in a machine-learning (ML) based paradigm. A database with 114 (28 high-risk) patients acquired from Lung Image Database Consortium (LIDC) is used in this study. After nodule segmentation using K-means clustering, features based on shape and texture attributes are extracted. Seven different filter and wrapper-based feature selection techniques are used for dominant feature selection. Lastly, the classification of nodules is performed by a support vector machine using six different kernel functions. The classification results are evaluated using 10-fold cross-validation and hold-out data division protocols. The performance of the proposed system is evaluated using accuracy, sensitivity, specificity, and the area under receiver operating characteristics (AUC). Using 30 dominant features from the pool of shape and texture-based features, the proposed system achieves the highest classification accuracy and AUC of 89% and 0.92, respectively. The proposed ML-based system showed an improvement in risk stratification accuracy by fusing shape and texture-based features.

show abstract

Section: Benchmarkingmentioning

confidence: 99%

Section: Benchmarkingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Improved pulmonary lung nodules risk stratification in computed tomography images by fusing shape and texture features in a machine‐learning paradigm

Sahu

Londhe

Verma

et al. 2020

Int J Imaging Syst Tech

View full text Add to dashboard Cite

show abstract

“…Lima et al [22] utilizou CNN 2D para classificação de nódulos pulmonares precoces (com diâmetro entre 5 e 10mm) em benignos e malignos. Os autores avaliaram diferentes estratégias classificar de um nódulo com base na classificação de cada um de seus cortes.…”

Section: Trabalhos Relacionadosunclassified

Pulmonary Nodule Classification with 3D Convolutional Neural Networks

Jatobá

Lima

Oliveira

2019

Anais Do XV Workshop De Visão Computacional (WVC 2019)

Self Cite

View full text Add to dashboard Cite

Lung cancer is a leading cause of death worldwide and its early detection is critical for patient survival. However, the diagnosis is still a challenging task, in which computeraided diagnosis (CADx) systems try to assist by providing a second opinion to a radiologist. In this work, we propose a 3D Convolutional Neural Network for classiﬁcation of solid pulmonary nodules into benign and malignant. We evaluated different approaches for the nodule volume assembling and tuned our models in an automated fashion. Our models achieved satisfactory results, with AUC of 0.89, accuracy of 81.37% and a sensibility of 84.83%. Moreover, our results have shown that the ﬁrst slices of a nodule provide the best results and only ﬁve nodule slices are enough for a 3D CNN achieve its best results.

show abstract

Hyperparameter for Deep Learning Applied in Mammogram Image Classification

Pereira

Ribeiro

2022

2022 IEEE 35th International Symposium on Computer-Based Medical Systems (CBMS)

View full text Add to dashboard Cite

Efficient Hyperparameter Optimization of Convolutional Neural Networks on Classification of Early Pulmonary Nodules

Cited by 6 publications

References 20 publications

Improved pulmonary lung nodules risk stratification in computed tomography images by fusing shape and texture features in a machine‐learning paradigm

Improved pulmonary lung nodules risk stratification in computed tomography images by fusing shape and texture features in a machine‐learning paradigm

Pulmonary Nodule Classification with 3D Convolutional Neural Networks

Hyperparameter for Deep Learning Applied in Mammogram Image Classification

Contact Info

Product

Resources

About