Deep convolutional neural network based medical image classification for disease diagnosis

Yadav, Samir S.; Jadhav, Shivajirao M.

doi:10.1186/s40537-019-0276-2

Cited by 807 publications

(392 citation statements)

References 23 publications

Supporting

Mentioning

384

Contrasting

Unclassified

Order By: Relevance

“…Landmark studies have tended to use a version of GoogleNet called Inception v3; however, AlexNet or other simple models such as VGG are popular in the analysis of medical data [ 23 ]. Indeed, VGG-16 has been reported to outperform Inception v3 in classifying medical images [ 24 , 25 ]. It has also accurately identified speckle patterns for the classification of hepatic steatosis [ 26 ].…”

Section: Introductionmentioning

confidence: 99%

Clinical Value of Information Entropy Compared with Deep Learning for Ultrasound Grading of Hepatic Steatosis

Chen

Chao

Tsai

et al. 2020

Entropy

View full text Add to dashboard Cite

Entropy is a quantitative measure of signal uncertainty and has been widely applied to ultrasound tissue characterization. Ultrasound assessment of hepatic steatosis typically involves a backscattered statistical analysis of signals based on information entropy. Deep learning extracts features for classification without any physical assumptions or considerations in acoustics. In this study, we assessed clinical values of information entropy and deep learning in the grading of hepatic steatosis. A total of 205 participants underwent ultrasound examinations. The image raw data were used for Shannon entropy imaging and for training and testing by the pretrained VGG-16 model, which has been employed for medical data analysis. The entropy imaging and VGG-16 model predictions were compared with histological examinations. The diagnostic performances in grading hepatic steatosis were evaluated using receiver operating characteristic (ROC) curve analysis and the DeLong test. The areas under the ROC curves when using the VGG-16 model to grade mild, moderate, and severe hepatic steatosis were 0.71, 0.75, and 0.88, respectively; those for entropy imaging were 0.68, 0.85, and 0.9, respectively. Ultrasound entropy, which varies with fatty infiltration in the liver, outperformed VGG-16 in identifying participants with moderate or severe hepatic steatosis (p < 0.05). The results indicated that physics-based information entropy for backscattering statistics analysis can be recommended for ultrasound diagnosis of hepatic steatosis, providing not only improved performance in grading but also clinical interpretations of hepatic steatosis.

show abstract

Section: Introductionmentioning

confidence: 99%

Clinical Value of Information Entropy Compared with Deep Learning for Ultrasound Grading of Hepatic Steatosis

Chen

Chao

Tsai

et al. 2020

Entropy

View full text Add to dashboard Cite

show abstract

“…In general VGG16 gets good results in many cases for finetuning task [30], this is the mean reason of choosing this model at first for transfer learning to process pneumonia classification.…”

Section: Validation and Testing The Given Cnn Modelsmentioning

confidence: 99%

Convolutional Neural Network Based Classification of Patients with Pneumonia using X-ray Lung Images

Moujahid¹,

Cherradi²,

Gannour³

et al. 2020

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

View full text Add to dashboard Cite

Analysis and classification of lung diseases using X-ray images is a primary step in the procedure of pneumonia diagnosis, especially in a critical period as pandemic of COVID-19 that is type of pneumonia. Therefore, an automatic method with high accuracy of classification is needed to perform classification of lung diseases due to the increasing number of cases. Convolutional Neural Networks (CNN) based classification has gained a big popularity over the last few years because of its speed and level of accuracy on the image's classification tasks. Through this article, we propose an implementation a CNNbased classification models using transfer learning technique to perform pneumonia detection and compare the results in order to detect the best model for the task according to certain parameters. As this has become a fast expanding field, there are several models but we will focus on the best outperforming algorithms according to their architecture, length and type of layers and evaluation parameters for the classification tasks. Firstly, we review the existing conventional methods and deep learning architectures used for segmentation in general. Next, we perform a deep performance and analysis based on accuracy and loss function of implemented models. A critical analysis of the results is made to highlight all important issues to improve.

show abstract

“…Shen et al [17] designed a method of multi-crop pooling that is used in DCNN to hold the object salient data which is applicable to classify the lung cancer under the application of CT images. Several other researchers have used DNN for image based disease classification [18][19][20][21][22][23]. Though various DCNN models have been presented in the literature, it is needed to increase the learning rate of DCNN.…”

Section: Introductionmentioning

confidence: 99%

Deep neural network with moth search optimization algorithm based detection and classification of diabetic retinopathy images

2020

View full text Add to dashboard Cite

In recent days, progressive rate of diabetic retinopathy (DR) becomes high and it is needed to develop an automated model for effective diagnosis of DR. This paper presents a new deep neural network with moth search optimization (DNN-MSO) algorithm based detection and classification model for DR images. The presented DNN-MSO algorithm involves different processes namely preprocessing, segmentation, feature extraction and classification. Initially, the contrast level of the DR images is improved using contrast limited adaptive histogram equalization model. After that, the preprocessed images are segmented using histogram approach. Then, Inception-ResNet V2 model is applied for feature extraction. Finally, extracted feature vectors are given to the DNN-MSO based classifier model to classify the different stages of DR. An extensive series of experiments were carried out and the results are validated on Messidor DR dataset. The obtained experimental outcome stated the superior characteristics of the DNN-MSO model by attaining a maximum accuracy, sensitivity and specificity of 99.12%, 97.91% and 99.47% respectively.

show abstract

Deep convolutional neural network based medical image classification for disease diagnosis

Cited by 807 publications

References 23 publications

Clinical Value of Information Entropy Compared with Deep Learning for Ultrasound Grading of Hepatic Steatosis

Clinical Value of Information Entropy Compared with Deep Learning for Ultrasound Grading of Hepatic Steatosis

Convolutional Neural Network Based Classification of Patients with Pneumonia using X-ray Lung Images

Deep neural network with moth search optimization algorithm based detection and classification of diabetic retinopathy images

Contact Info

Product

Resources

About