Convolutional neural networks in medical image understanding: a survey

Sarvamangala, D. R.; Kulkarni, Raghavendra V.

doi:10.1007/s12065-020-00540-3

Cited by 434 publications

(246 citation statements)

References 64 publications

Supporting

Mentioning

244

Contrasting

Unclassified

Order By: Relevance

“…Due to their high predictive power, neural networks are extensively used in biomedical image classification tasks. Sarvamangala surveys CNNs for medical image understanding [ 6 ]. Litjens et al summarize 300 papers on deep learning for medical image analysis [ 7 ].…”

Section: Related Workmentioning

confidence: 99%

Evaluating Deep Neural Network Architectures with Transfer Learning for Pneumonitis Diagnosis

Krishnamurthy¹,

Srinivasan

Qaisar

et al. 2021

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

Pneumonitis is an infectious disease that causes the inflammation of the air sac. It can be life-threatening to the very young and elderly. Detection of pneumonitis from X-ray images is a significant challenge. Early detection and assistance with diagnosis can be crucial. Recent developments in the field of deep learning have significantly improved their performance in medical image analysis. The superior predictive performance of the deep learning methods makes them ideal for pneumonitis classification from chest X-ray images. However, training deep learning models can be cumbersome and resource-intensive. Reusing knowledge representations of public models trained on large-scale datasets through transfer learning can help alleviate these challenges. In this paper, we compare various image classification models based on transfer learning with well-known deep learning architectures. The Kaggle chest X-ray dataset was used to evaluate and compare our models. We apply basic data augmentation and fine-tune our feed-forward classification head on the models pretrained on the ImageNet dataset. We observed that the DenseNet201 model outperforms other models with an AUROC score of 0.966 and a recall score of 0.99. We also visualize the class activation maps from the DenseNet201 model to interpret the patterns recognized by the model for prediction.

show abstract

Section: Related Workmentioning

confidence: 99%

Evaluating Deep Neural Network Architectures with Transfer Learning for Pneumonitis Diagnosis

Krishnamurthy¹,

Srinivasan

Qaisar

et al. 2021

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

show abstract

“…Among deep learning algorithms, autoencoders have been increasingly used in medical applications in recent years 26 , 27 . Although CNN is one of the most known and used deep learning algorithms, it is often preferred especially in the analysis of image data and provides successful results 28 , 29 . Therefore, to classify 1D SERS signals of MRSA and MSSA stacked autoencoder (SAE) based deep neural network (DNN) was preferred in this study.…”

Section: Introductionmentioning

confidence: 99%

Drug-resistant Staphylococcus aureus bacteria detection by combining surface-enhanced Raman spectroscopy (SERS) and deep learning techniques

Ciloglu

Çalışkan

Saridag

et al. 2021

Sci Rep

View full text Add to dashboard Cite

Over the past year, the world's attention has focused on combating COVID-19 disease, but the other threat waiting at the door—antimicrobial resistance should not be forgotten. Although making the diagnosis rapidly and accurately is crucial in preventing antibiotic resistance development, bacterial identification techniques include some challenging processes. To address this challenge, we proposed a deep neural network (DNN) that can discriminate antibiotic-resistant bacteria using surface-enhanced Raman spectroscopy (SERS). Stacked autoencoder (SAE)-based DNN was used for the rapid identification of methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-sensitive S. aureus (MSSA) bacteria using a label-free SERS technique. The performance of the DNN was compared with traditional classifiers. Since the SERS technique provides high signal-to-noise ratio (SNR) data, some subtle differences were found between MRSA and MSSA in relative band intensities. SAE-based DNN can learn features from raw data and classify them with an accuracy of 97.66%. Moreover, the model discriminates bacteria with an area under curve (AUC) of 0.99. Compared to traditional classifiers, SAE-based DNN was found superior in accuracy and AUC values. The obtained results are also supported by statistical analysis. These results demonstrate that deep learning has great potential to characterize and detect antibiotic-resistant bacteria by using SERS spectral data.

show abstract

“…According to the review conducted in [ 44 ], many applications surpass or equal the results obtained by human specialists. Although the use of DL has advanced in the medical field, including CAD [ 45 , 46 , 47 ], the application of these techniques in US images can be considered incipient [ 48 ]. Considering the main global issue since the last year, the COVID-19 pandemic, there are works on DL related with promising results, which corroborates the clinical findings in the area.…”

Section: Introductionmentioning

confidence: 99%

Pulmonary COVID-19: Learning Spatiotemporal Features Combining CNN and LSTM Networks for Lung Ultrasound Video Classification

Barros

Lacerda

Albuquerque

et al. 2021

Sensors

View full text Add to dashboard Cite

Deep Learning is a very active and important area for building Computer-Aided Diagnosis (CAD) applications. This work aims to present a hybrid model to classify lung ultrasound (LUS) videos captured by convex transducers to diagnose COVID-19. A Convolutional Neural Network (CNN) performed the extraction of spatial features, and the temporal dependence was learned using a Long Short-Term Memory (LSTM). Different types of convolutional architectures were used for feature extraction. The hybrid model (CNN-LSTM) hyperparameters were optimized using the Optuna framework. The best hybrid model was composed of an Xception pre-trained on ImageNet and an LSTM containing 512 units, configured with a dropout rate of 0.4, two fully connected layers containing 1024 neurons each, and a sequence of 20 frames in the input layer (20×2018). The model presented an average accuracy of 93% and sensitivity of 97% for COVID-19, outperforming models based purely on spatial approaches. Furthermore, feature extraction using transfer learning with models pre-trained on ImageNet provided comparable results to models pre-trained on LUS images. The results corroborate with other studies showing that this model for LUS classification can be an important tool in the fight against COVID-19 and other lung diseases.

show abstract

Convolutional neural networks in medical image understanding: a survey

Cited by 434 publications

References 64 publications

Evaluating Deep Neural Network Architectures with Transfer Learning for Pneumonitis Diagnosis

Evaluating Deep Neural Network Architectures with Transfer Learning for Pneumonitis Diagnosis

Drug-resistant Staphylococcus aureus bacteria detection by combining surface-enhanced Raman spectroscopy (SERS) and deep learning techniques

Pulmonary COVID-19: Learning Spatiotemporal Features Combining CNN and LSTM Networks for Lung Ultrasound Video Classification

Contact Info

Product

Resources

About