SpinalNet: Deep Neural Network With Gradual Input

Kabir, H M Dipu; Abdar, Moloud; Jalali, Seyed Mohammad Jafar; Khosravi, Abbas; Atiya, Amir F.; Nahavandi, Saeid; Srinivasan, Dipti

doi:10.1109/tai.2022.3185179

Cited by 80 publications

(37 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this research work, the author used wellknown MLA approaches to examine actual diagnostic medical data based on various risk factors to assess their effectiveness for diabetic probability. Seven MLA were used in this study such as RF, KNN, MLP, SVC, GBC, DT, and LR [24][25][26][27][28][29][30][31][32]. Various statistical criteria were used to compare the analytical results.…”

Section: Literature Reviewmentioning

confidence: 99%

Prevalence and Early Prediction of Diabetes Using Machine Learning in North Kashmir: A Case Study of District Bandipora

Bhat

Selvam

Ansari

et al. 2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Diabetes is one of the biggest health problems that affect millions of people across the world. Uncontrolled diabetes can increase the risk of heart attack, cancer, kidney damage, blindness, and other illnesses. Researchers are motivated to create a Machine Learning methodology that can predict diabetes in the future. Exploiting Machine Learning Algorithms (MLA) is essential if healthcare professionals are able to identify diseases more effectively. In order to improve the medical diagnosis of diabetes this research explored and contrasts various MLA that can identify diabetes risk early. The research includes the analysis on real datasets such as a clinical dataset gathered from a doctor in the Indian district of Bandipora in the years April 2021–Feb2022. MLA are currently important in the healthcare sector due to their prediction abilities. Researchers are using MLA to improve disease prediction and reduce cost. In this Paper author developed a methodology using Machine Learning Algorithms for Diabetes Disease Risk Prediction in North Kashmir. Six MLA have been successfully used in the experimental study such as Random Forest (RF), Multi-Layer Perceptron (MLP), Support Vector Machine (SVM), Gradient Boost (GB), Decision Tree (DT), and Logistic Regression (LR). RF is the most accurate classifier with the uppermost accuracy rate of 98 percent followed by MLP (90.99%), SVM (92%), GBC (97%), DT (96%), and LR (69%), respectively, with the balanced data set. Lastly, this study enables us to effectively identify the prevalence and prediction of diabetes.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Prevalence and Early Prediction of Diabetes Using Machine Learning in North Kashmir: A Case Study of District Bandipora

Bhat

Selvam

Ansari

et al. 2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…To investigate spatial and spectral properties, the model uses a combination of 3D and 2D convolution layers. The HybridMSSN deep learning model uses five Spinal Fully Connected Networks (SFCNs) or SpinalNets (Kabir et al, 2020) to improve classification accuracy even with minimal training samples. where N is the total number of labeled pixels of the given image, and s is the patch size.…”

Section: Proposed Frameworkmentioning

confidence: 99%

“…In the proposed model the dense layer is replaced by Spinal Fully Connected Network (SFCN) or SpinalNet (Kabir et al, 2020) 2.2.4 Spinal Fully Connected Layer (SFCN) or SpinalNet The SFCN (Kabir et al, 2020) aims to replicate the human somatosensory system. Each layer of the SFCN contributes to the local output, which is comparable to the reflex system, and a portion of the modified input is delivered to the global output, that is analogous to the brain.…”

Section: Multiscale 3d-2d-cnnmentioning

confidence: 99%

“…Two extremely important preset parameters in the SFCN structure are layer width and Half width. Layer width is set to an appropriate random value equal to the number of filters used in the 2D convolution layer, and half width is set to half of the layer width rounded off to an integer value based on the literature discovered in (Kabir et al, 2020). The SFCN substitutes four sublayers for the hidden layers, each of which has an input, dense, dropout, and concatenate network.…”

Section: Multiscale 3d-2d-cnnmentioning

confidence: 99%

See 1 more Smart Citation

HybridMSSN: A Deep Learning Hyperspectral Image Classification Model

Ladi

PANDA²,

Dash

et al. 2022

Preprint

View full text Add to dashboard Cite

With the goal of obtaining accuracy and optimum computation, Hyperspectral image (HI) classification models have been developed for best use in a variety of applications. Because HI contains a lot of spatial and spectral information, classification using merely a Convolutional Neural Network (CNN) isn't always enough. This paper solves these concerns by proposing a new Hybrid Multi-Scale Spinal Net (HybridMSSN). The HI is subjected to Principal Component Analysis (PCA) in order to isolate the most useful spectral bands and eliminate the superfluous ones. For feature learning and classification, the model uses multiscale CNNs and a Spinal Fully Connected Network (SFCN). Three 3D-CNNs are used to extract spectral and spatial features from the 3-dimensional HI, followed by a 2D-CNN for spatial feature learning. The model uses SFCNs to optimize computation, reducing the number of multiplications in the activation function of neural networks and resulting in a faster response. The multi-scale features acquired from the different 3D Convolution layers are flattened and transferred to independent SFCNs after Max Pooling, Dropout, and Batch Normalization. Similarly, features collected from the 2D convolution layer and Dropout are flattened and supplied to the SFCN. For categorization, the SFCNs' outputs are combined. Despite limited training samples, the existence of noise, and an imbalance of class concerns, investigations conducted on three standard datasets revealed significant accuracy gains when compared to four state-of-the-art models.

show abstract

“…The development of computer vision has experienced a range of trends in this decade. Several introducing models [ 1 , 2 , 3 , 4 , 5 , 6 ] in open datasets competition significantly improved the accuracy of the image classification, which includes deep convolutional neural networks (CNNs) with residual calculation [ 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 ]. With the deep CNNs from stacking convolutional layers, models could capture local characteristics and global profiles with the increasing receptive fields [ 15 ].…”

Section: Introductionmentioning

confidence: 99%

UPANets: Learning from the Universal Pixel Attention Neworks

Tseng

Lee

Feng

et al. 2022

Entropy

View full text Add to dashboard Cite

With the successful development in computer vision, building a deep convolutional neural network (CNNs) has been mainstream, considering the character of shared parameters in a convolutional layer. Stacking convolutional layers into a deep structure improves performance, but over-stacking also ramps up the needed resources for GPUs. Seeing another surge of Transformers in computer vision, the issue has aroused severely. A resource-hungry model is hardly implemented for limited hardware or single-customers-based GPU. Therefore, this work focuses on these concerns and proposes an efficient but robust backbone, which equips with channel and spatial direction attentions, so the attentions help to expand receptive fields in shallow convolutional layers and pass the information to every layer. An attention-boosted network based on already efficient CNNs, Universal Pixel Attention Networks (UPANets), is proposed. Through a series of experiments, UPANets fulfil the purposes of learning global information with less needed resources and outshine many existing SOTAs in CIFAR-{10, 100}.

show abstract

SpinalNet: Deep Neural Network With Gradual Input

Cited by 80 publications

References 52 publications

Prevalence and Early Prediction of Diabetes Using Machine Learning in North Kashmir: A Case Study of District Bandipora

Prevalence and Early Prediction of Diabetes Using Machine Learning in North Kashmir: A Case Study of District Bandipora

HybridMSSN: A Deep Learning Hyperspectral Image Classification Model

UPANets: Learning from the Universal Pixel Attention Neworks

Contact Info

Product

Resources

About