Optimized Deep Convolutional Neural Networks for Identification of Macular Diseases from Optical Coherence Tomography Images

Ji, Qingge; Huang, Jie; We, He; Sun, Yankui

doi:10.3390/a12030051

Cited by 76 publications

(20 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Fig 6(A)–6(E) demonstrate how factorization of 5×5, 7×7 convolutions into smaller convolutions 3×3, or asymmetric convolutions 1×7 and 7×1, is conducted during the experiment. This reduces the number of deep network parameters [ 46 ]. Compared against the new classification layers added, the original classification layers replaced include the global average pooling layer and the last dense layer with 1000 outputs.…”

Section: Methodsmentioning

confidence: 99%

“…In the case of a convolution block, its input dimensions are smaller than its output dimensions due to the convolutional layer’s availability at a shortcut ( Fig 8(C) ). In both blocks, 1×1 convolution is implemented at the beginning, and the end of the network through a technique called a bottleneck design [ 46 ]. This technique decreases the number of parameters without degrading the performance of the network.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Gaussian smoothing and modified histogram normalization methods to improve neural-biomarker interpretations for dyslexia classification mechanism

et al. 2021

View full text Add to dashboard Cite

Achieving biologically interpretable neural-biomarkers and features from neuroimaging datasets is a challenging task in an MRI-based dyslexia study. This challenge becomes more pronounced when the needed MRI datasets are collected from multiple heterogeneous sources with inconsistent scanner settings. This study presents a method of improving the biological interpretation of dyslexia’s neural-biomarkers from MRI datasets sourced from publicly available open databases. The proposed system utilized a modified histogram normalization (MHN) method to improve dyslexia neural-biomarker interpretations by mapping the pixels’ intensities of low-quality input neuroimages to range between the low-intensity region of interest (ROIlow) and high-intensity region of interest (ROIhigh) of the high-quality image. This was achieved after initial image smoothing using the Gaussian filter method with an isotropic kernel of size 4mm. The performance of the proposed smoothing and normalization methods was evaluated based on three image post-processing experiments: ROI segmentation, gray matter (GM) tissues volume estimations, and deep learning (DL) classifications using Computational Anatomy Toolbox (CAT12) and pre-trained models in a MATLAB working environment. The three experiments were preceded by some pre-processing tasks such as image resizing, labelling, patching, and non-rigid registration. Our results showed that the best smoothing was achieved at a scale value, σ = 1.25 with a 0.9% increment in the peak-signal-to-noise ratio (PSNR). Results from the three image post-processing experiments confirmed the efficacy of the proposed methods. Evidence emanating from our analysis showed that using the proposed MHN and Gaussian smoothing methods can improve comparability of image features and neural-biomarkers of dyslexia with a statistically significantly high disc similarity coefficient (DSC) index, low mean square error (MSE), and improved tissue volume estimations. After 10 repeated 10-fold cross-validation, the highest accuracy achieved by DL models is 94.7% at a 95% confidence interval (CI) level. Finally, our finding confirmed that the proposed MHN method significantly outperformed the normalization method of the state-of-the-art histogram matching.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Gaussian smoothing and modified histogram normalization methods to improve neural-biomarker interpretations for dyslexia classification mechanism

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The identity block has not convolution layer in shortcut, and the output possesses the same dimension as the input dimension. [ 28 ] As shown in Figure 3 both these blocks consist of two 3 × 3 convolution layers, followed by ReLU and Batch normalization layers.…”

Section: Proposed Methodsmentioning

confidence: 99%

Residual learning: A new paradigm to improve deep learning-based segmentation of the left ventricle in magnetic resonance imaging cardiac images

et al. 2021

View full text Add to dashboard Cite

Background: Recently, magnetic resonance imaging (MRI) has become a useful tool for the early detection of heart failure. A vital step of this process is a valid measurement of the left ventricle's properties, which seriously depends on the accurate segmentation of the heart in captured images. Although various schemes have been tested for this segmentation so far, the latest proposed methods have used the concept of deep learning to estimate the range of the left ventricle in cardiac MRI images. While deep learning methods can lead to better results than their classical alternatives, but unfortunately, the gradient vanishing and exploding problems may hamper their efficiency for the accurate segmentation of the left ventricle in MRI heart images. Methods: In this article, a new concept called residual learning is utilized to improve the performance of deep learning schemes against gradient vanishing problems. For this purpose, the Residual Network of Residual Network (i.e., Residual of Residual) substructure is utilized inside the main deep learning architecture (e.g., Unet), which provides more significant detection indexes. Results and Conclusion: The proposed method's performances and its alternatives were evaluated on Sunnybrook Cardiac Data as a reliable dataset in the left ventricle segmentation. The results show that the detection parameters are improved at least by 5%, 3.5%, 8.1%, and 11.4% compared to its deep alternatives in terms of Jaccard, Dice, precision, and false-positive rate indexes, respectively. These improvements were made when the recall parameter was reduced to a negligible value (i.e., approximately 1%). Overall, the proposed method can be used as a suitable tool for more accurate detection of the left ventricle in MRI images.

show abstract

“…Once the three layers are added to the implemented CNN architectures, we use "Adam" as the adaptive optimizer of the learning rate. Figure 5 displays the summary of DenseNet121 architecture with the three added layers, and Figure 6 shows the detailed DenseNet121 architecture [39]. We have built the eight CNN architectures that we have already mentioned in Int J Artif Intell ISSN: 2252-8938…”

Section: Network Building and Learningmentioning

confidence: 99%

Deep convolutional neural networks architecture for an efficient emergency vehicle classification in real-time traffic monitoring

Kherraki

Ouazzani

2022

IJ-AI

View full text Add to dashboard Cite

Nowadays, intelligent transportation system (ITS) has become one of the most popular subjects of scientific research. ITS provides innovative services to traffic monitoring. The classification of emergency vehicles in traffic surveillance cameras provides an early warning to ensure a rapid reaction in emergency events. Computer vision technology, including deep learning, has many advantages for traffic monitoring. For instance, convolutional neural network (CNN) has given very good results and optimal performance in computer vision tasks, such as the classification of vehicles according to their types, and brands. In this paper, we will classify emergency vehicles from the output of a closed-circuit television (CCTV) camera. Among the advantages of this research paper is providing detailed information on the emergency vehicle classification topic. Emergency vehicles have the highest priority on the road and finding the best emergency vehicle classification model in realtime will undoubtedly save lives. Thus, we have used eight CNN architectures and compared their performances on the Analytics Vidhya Emergency Vehicle dataset. The experiments show that the utilization of DenseNet121 gives excellent classification results which makes it the most suitable architecture for this research topic, besides, DenseNet121 does not require a high memory size which makes it appropriate for real-time applications.<p> </p>

show abstract

Optimized Deep Convolutional Neural Networks for Identification of Macular Diseases from Optical Coherence Tomography Images

Cited by 76 publications

References 30 publications

Gaussian smoothing and modified histogram normalization methods to improve neural-biomarker interpretations for dyslexia classification mechanism

Gaussian smoothing and modified histogram normalization methods to improve neural-biomarker interpretations for dyslexia classification mechanism

Residual learning: A new paradigm to improve deep learning-based segmentation of the left ventricle in magnetic resonance imaging cardiac images

Deep convolutional neural networks architecture for an efficient emergency vehicle classification in real-time traffic monitoring

Contact Info

Product

Resources

About