OctNET: A Lightweight CNN for Retinal Disease Classification from Optical Coherence Tomography Images.

Abstract:

BACKGROUND AND OBJECTIVE:Retinal diseases are becoming a major health problem in recent years. Their early detection and ensuing treatment are essential to prevent visual damage, as the number of people affected by diabetes is expected to grow exponentially. Retinal diseases progress slowly, without any discernible symptoms. Optical Coherence Tomography (OCT) is a diagnostic tool capable of analyzing and identifying the quantitative discrimination in the disease affected retinal layers with high resolution. This paper proposes a deep neural network-based classifier for the computer-aided classification of Diabetic Macular Edema (DME), drusen, Choroidal NeoVascularization (CNV) from normal OCT images of the retina. METHODS:In the proposed method, we demonstrate the feasibility of classifying and detecting severe retinal pathologies from OCT images using a deep convolutional neural network having six convolutional blocks. The classification results are explained using a gradient-based class activation mapping algorithm. RESULTS:Training and validation of the model are performed on a public dataset of 83,484 images with expert-level disease grading of CNV, DME, and drusen, in addition to normal retinal image. We achieved a precision of 99.69%, recall of 99.69%, and accuracy of 99.69% with only three misclassifications out of 968 test cases. CONCLUSION:In the proposed work, downsampling and weight sharing were introduced to improve the training efficiency and were found to reduce the trainable parameters significantly. The class activation mapping was also performed, and the output image was similar to the retina's actual color OCT images. The proposed network used only 6.9% of learnable parameters compared to the existing ResNet-50 model and yet outperformed it in classification. The proposed work can be potentially employed in real-time applications due to reduced complexity and fewer learnable parameters over other models.

authors

A P S,Kar S,S G,Gopi VP,Palanisamy P

doi

10.1016/j.cmpb.2020.105877

subject

Has Abstract

pub_date

2020-11-28 00:00:00

pages

105877

eissn

0169-2607

issn

1872-7565

pii

S0169-2607(20)31710-7

pub_type

杂志文章
  • Cascaded classifiers and stacking methods for classification of pulmonary nodule characteristics.

    abstract:BACKGROUND AND OBJECTIVES:Detection and classification of pulmonary nodules are critical tasks in medical image analysis. The Lung Image Database Consortium (LIDC) database is a widely used resource for small pulmonary nodule classification research. This dataset is comprised of nodule characteristic evaluations and CT...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2018.10.009

    authors: Kaya A

    更新日期:2018-11-01 00:00:00

  • Two-step deep neural network for segmentation of deep white matter hyperintensities in migraineurs.

    abstract:BACKGROUND AND OBJECTIVE:Patients with migraine show an increased presence of white matter hyperintensities (WMHs), especially deep WMHs. Segmentation of small, deep WMHs is a critical issue in managing migraine care. Here, we aim to develop a novel approach to segmenting deep WMHs using deep neural networks based on t...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2019.105065

    authors: Hong J,Park BY,Lee MJ,Chung CS,Cha J,Park H

    更新日期:2020-01-01 00:00:00

  • Microcomputer program for automated action potential waveform analysis.

    abstract::A program for action potential waveform analysis based on a PC compatible computer is described. Single or averaged action potentials are analyzed by obtaining its first derivative and using criteria which allow automatic measurement of several action potential components, including: depolarization rate, repolarizatio...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/s0169-2607(00)00053-5

    authors: Soto E,Salceda E,Cruz R,Ortega A,Vega R

    更新日期:2000-06-01 00:00:00

  • Volumetric interpolation of tomographic sequences for accurate 3D reconstruction of anatomical parts.

    abstract:BACKGROUND AND OBJECTIVE:Tomographic sequences of biomedical images are commonly used to achieve a three-dimensional visualization of the human anatomy. In some cases, the number of images contained in the sequence is limited, e.g., in low-dose computed tomography acquired on neonatal patients, resulting in a coarse an...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2020.105525

    authors: Santarelli C,Argenti F,Uccheddu F,Alparone L,Carfagni M

    更新日期:2020-10-01 00:00:00

  • BE-DTI': Ensemble framework for drug target interaction prediction using dimensionality reduction and active learning.

    abstract:BACKGROUND AND OBJECTIVE:Drug-target interaction prediction plays an intrinsic role in the drug discovery process. Prediction of novel drugs and targets helps in identifying optimal drug therapies for various stringent diseases. Computational prediction of drug-target interactions can help to identify potential drug-ta...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2018.08.011

    authors: Sharma A,Rani R

    更新日期:2018-10-01 00:00:00

  • The stable status evaluation for female breast implant surgery by calculating related physics parameters.

    abstract::Cosmetic doctor utilizes the position, size and shapes of female's breast to judge whether the breast is under steady-state condition after breast implant plastic surgery. Since, doctor evaluates the breast condition with the subjective discrimination (such as vision, sense of touch) without using the objective physic...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2007.11.013

    authors: Sun SP,Hsu KW,Chen JS

    更新日期:2008-05-01 00:00:00

  • Global parenchymal texture features based on histograms of oriented gradients improve cancer development risk estimation from healthy breasts.

    abstract:BACKGROUND:The breast dense tissue percentage on digital mammograms is one of the most commonly used markers for breast cancer risk estimation. Geometric features of dense tissue over the breast and the presence of texture structures contained in sliding windows that scan the mammograms may improve the predictive abili...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2019.05.022

    authors: Pérez-Benito FJ,Signol F,Pérez-Cortés JC,Pollán M,Pérez-Gómez B,Salas-Trejo D,Casals M,Martínez I,LLobet R

    更新日期:2019-08-01 00:00:00

  • MODDIS: a microcomputer program for model discrimination.

    abstract::A computer program for a microcomputer (HP 86) is presented to discriminate between different models and to design new experiments for model discrimination. By a non-linear fitting algorithm a set of experimental data is fitted with different models suggested by the user. The parameters characterizing each model are e...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/0169-2607(86)90022-2

    authors: Verotta D,Recchia M,Urso R

    更新日期:1986-04-01 00:00:00

  • Multi-atlas based segmentation using probabilistic label fusion with adaptive weighting of image similarity measures.

    abstract::Label fusion multi-atlas approaches for image segmentation can give better segmentation results than single atlas methods. We present a multi-atlas label fusion strategy based on probabilistic weighting of distance maps. Relationships between image similarities and segmentation similarities are estimated in a learning...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2012.12.006

    authors: Sjöberg C,Ahnesjö A

    更新日期:2013-06-01 00:00:00

  • Improving drug dosing in hospitalized patients: automated modeling of pharmacokinetics for individualization of drug dosage regimens.

    abstract::Many clinical useful drugs have a narrow range of blood concentrations which are both safe and efficacious. Inter-individual and intra-individual variations in drug disposition are important factors causing blood concentrations of drug to fall outside of the therapeutic range. Modeling of the pharmacokinetics of the i...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/0169-2607(89)90069-2

    authors: Lenert L,Sheiner L,Blaschke T

    更新日期:1989-10-01 00:00:00

  • Single subject image analysis using the complex general linear model--an application to functional magnetic resonance imaging with multiple inputs.

    abstract::A linear time invariant model is applied to functional fMRI blood flow data. Based on traditional time series analysis, this model assumes that the fMRI stochastic output sequence can be determined by a constant plus a linear filter (hemodynamic response function) of several fixed deterministic inputs and an error ter...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2005.12.003

    authors: Rio DE,Rawlings RR,Woltz LA,Salloum JB,Hommer DW

    更新日期:2006-04-01 00:00:00

  • Interstitial insulin kinetic parameters for a 2-compartment insulin model with saturable clearance.

    abstract::Glucose-insulin system models are commonly used for identifying insulin sensitivity. With physiological, 2-compartment insulin kinetics models, accurate kinetic parameter values are required for reliable estimates of insulin sensitivity. This study uses data from 6 published microdialysis studies to determine the most...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2014.01.011

    authors: Pretty CG,Le Compte A,Penning S,Fisk L,Shaw GM,Desaive T,Chase JG

    更新日期:2014-05-01 00:00:00

  • Computer-based assessment for facioscapulohumeral dystrophy diagnosis.

    abstract::The paper presents a computer-based assessment for facioscapulohumeral dystrophy (FSHD) diagnosis through characterisation of the fat and oedema percentages in the muscle region. A novel multi-slice method for the muscle-region segmentation in the T1-weighted magnetic resonance images is proposed using principles of t...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2015.03.006

    authors: Chambers O,Milenković J,Pražnikar A,Tasič JF

    更新日期:2015-06-01 00:00:00

  • Max-margin weight learning for medical knowledge network.

    abstract:BACKGROUND AND OBJECTIVE:The application of medical knowledge strongly affects the performance of intelligent diagnosis, and method of learning the weights of medical knowledge plays a substantial role in probabilistic graphical models (PGMs). The purpose of this study is to investigate a discriminative weight-learning...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2018.01.005

    authors: Jiang J,Xie J,Zhao C,Su J,Guan Y,Yu Q

    更新日期:2018-03-01 00:00:00

  • Semi-supervised learning for automatic segmentation of the knee from MRI with convolutional neural networks.

    abstract:BACKGROUND AND OBJECTIVE:Segmentation is a crucial step in multiple biomechanics and orthopedics applications. The time-intensiveness and expertise requirements of medical image segmentation present a significant bottleneck for corresponding workflows. The current study develops and evaluates convolutional neural netwo...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2020.105328

    authors: Burton W 2nd,Myers C,Rullkoetter P

    更新日期:2020-06-01 00:00:00

  • GPU-based parallel group ICA for functional magnetic resonance data.

    abstract::The goal of our study is to develop a fast parallel implementation of group independent component analysis (ICA) for functional magnetic resonance imaging (fMRI) data using graphics processing units (GPU). Though ICA has become a standard method to identify brain functional connectivity of the fMRI data, it is computa...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2015.02.002

    authors: Jing Y,Zeng W,Wang N,Ren T,Shi Y,Yin J,Xu Q

    更新日期:2015-04-01 00:00:00

  • Correlation method versus enhanced modified moving average method for automatic detection of T-wave alternans.

    abstract::Enhanced modified moving average method (EMMAM) and correlation method (CM) for microvolt TWA identification are compared by aid of simulated ECG tracings (cases of absence of TWA and presence of stationary or time-varying TWA) and ECG recordings from healthy subjects (H-group) and patients who survived an acute myoca...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2010.01.008

    authors: Burattini L,Bini S,Burattini R

    更新日期:2010-04-01 00:00:00

  • Feedback controlled electrical nerve stimulation: a computer simulation.

    abstract::The role of repetitive firing in neurophysiologic or neuropsychiatric disorders, such as Parkinson, epilepsy and bipolar type disorders, has always been a topic of medical research as therapies target either the cease of firing or a decrease in its frequency. In electrotherapy, one of the mechanisms to achieve the pur...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2010.01.006

    authors: Doruk RO

    更新日期:2010-07-01 00:00:00

  • Segmentation of prostate zones using probabilistic atlas-based method with diffusion-weighted MR images.

    abstract:BACKGROUND AND OBJECTIVE:Accurate segmentation of prostate and its zones constitute an essential preprocessing step for computer-aided diagnosis and detection system for prostate cancer (PCa) using diffusion-weighted imaging (DWI). However, low signal-to-noise ratio and high variability of prostate anatomic structures ...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2020.105572

    authors: Singh D,Kumar V,Das CJ,Singh A,Mehndiratta A

    更新日期:2020-11-01 00:00:00

  • Biomechanical effects of the working modes of LVADs on the aortic valve: A primary numerical study.

    abstract::Aortic valve diseases caused by the support from left ventricular assist devices (LVADs) have attracted increasing attention due to the wide application of the LVADs. However, the biomechanical effects of the working modes of LVADs on the aortic valve are still poorly understood. Hence, in this study, these biomechani...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2020.105512

    authors: Gao B,Zhang Q

    更新日期:2020-09-01 00:00:00

  • Unique parameter identification for cardiac diagnosis in critical care using minimal data sets.

    abstract::Lumped parameter approaches for modelling the cardiovascular system typically have many parameters of which a significant percentage are often not identifiable from limited data sets. Hence, significant parts of the model are required to be simulated with little overall effect on the accuracy of data fitting, as well ...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2010.01.002

    authors: Hann CE,Chase JG,Desaive T,Froissart CB,Revie J,Stevenson D,Lambermont B,Ghuysen A,Kolh P,Shaw GM

    更新日期:2010-07-01 00:00:00

  • Separation of P300 event-related potential using time varying time-lag blind source separation algorithm.

    abstract::Synchronous averaging over time locked single-trial of event-related potential (ERP) is known as the simplest scheme to extract P300 component. This method assumes the P300 features are invariant through the time while they are affected by factors like brain fatigue and habitation. In this study, a new scheme is propo...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2017.04.014

    authors: Sabeti M,Boostani R

    更新日期:2017-07-01 00:00:00

  • predicting and improving the probability of live birth for women undergoing frozen-thawed embryo transfer: a data-driven estimation and simulation model.

    abstract:BACKGROUND AND OBJECTIVE:Frozen-thawed embryo transfer (FET) is now widely used for the treatment of infertility. For many couples and clinicians, concerns over the probability and how to increase the chance of a successful birth are very common. Currently, there is not a single model to predict the live birth outcomes...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2020.105780

    authors: Liang R,An J,Zheng Y,Li J,Wang Y,Jia Y,Zhang J,Lu Q

    更新日期:2021-01-01 00:00:00

  • Capturing the semantic relationship between clinical terms with current MeSH bibliographic coding.

    abstract::This paper compares bibliographic retrieval using current MeSH (Medical Subject Headings) to bibliographic retrieval using explicitly coded semantic relationships between index terms. In a previous study, ten lists of abstracts, each list containing 20-40 papers discussing a specific pair of terms, were analyzed to id...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/0169-2607(88)90084-3

    authors: Miller PL,Smith P,Morrow JS,Riely CA,Powsner SM

    更新日期:1988-11-01 00:00:00

  • Finite element analysis of torque induced orthodontic bracket slot deformation in various bracket-archwire contact assembly.

    abstract:BACKGROUND AND OBJECTIVES:Orthodontic fixed appliance therapy involves alignment of teeth through the bracket and archwires. The archwire twist (torque) imparts significant forces inside the bracket slot in refining the teeth position at the end of treatment. The objective of this in- silico study was to evaluate the t...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2020.105748

    authors: Harikrishnan P,Magesh V,Ajayan AM,JebaSingh DK

    更新日期:2020-12-01 00:00:00

  • Design of crossover microcomputer program and application on drug bioequivalence data.

    abstract::A simple microcomputer crossover program in BASIC is designed for the purpose of drug bioequivalence studies and other related fields of statistical investigation. An illustration of the application of the program is presented to compare the bioavailability of ibuprofen when administered alone or followed by colestipo...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 临床试验,杂志文章

    doi:10.1016/0169-2607(95)01695-3

    authors: Abdullah ME,El-Sayed YM

    更新日期:1995-12-01 00:00:00

  • TAQIH, a tool for tabular data quality assessment and improvement in the context of health data.

    abstract:BACKGROUND AND OBJECTIVES:Data curation is a tedious task but of paramount relevance for data analytics and more specially in the health context where data-driven decisions must be extremely accurate. The ambition of TAQIH is to support non-technical users on 1) the exploratory data analysis (EDA) process of tabular he...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2018.12.029

    authors: Álvarez Sánchez R,Beristain Iraola A,Epelde Unanue G,Carlin P

    更新日期:2019-11-01 00:00:00

  • Analysis of sub-anatomic diffusion tensor imaging indices in white matter regions of Alzheimer with MMSE score.

    abstract::In this study, an attempt has been made to find the correlation between diffusion tensor imaging (DTI) indices of white matter (WM) regions and mini mental state examination (MMSE) score of Alzheimer patients. Diffusion weighted images are obtained from the ADNI database. These are preprocessed for eddy current correc...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2014.06.004

    authors: Patil RB,Ramakrishnan S

    更新日期:2014-10-01 00:00:00

  • Finite element analysis of immature teeth filled with MTA, Biodentine and Bioaggregate.

    abstract:BACKGROUND AND OBJECTIVE:Finite element based simulation has emerged as a powerful tool to analyse the tooth strength and its fracture characteristics. The aim of this study is to compare and evaluate the fracture resistance of immature teeth reinforcement with MTA, Biodentine and Bioaggregate as an apical plug and bac...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/j.cmpb.2020.105356

    authors: Eram A,Zuber M,Keni LG,Kalburgi S,Naik R,Bhandary S,Amin S,Badruddin IA

    更新日期:2020-07-01 00:00:00

  • Development and evaluation of PC-based HIS-RIS-modality-PACS coupling: the results of evaluation of initial stage with personal computer application.

    abstract::A PC-based HIS/RIS/Modality/PACS coupling was developed and operated where hospital-wide PACS was closely cooperated with the HIS and RIS. Three workstations with a magneto-optical disk (MOD) juke-box for each one were used as PACS servers to collect images from eight computed radiography systems and three CT scanners...

    journal_title:Computer methods and programs in biomedicine

    pub_type: 杂志文章

    doi:10.1016/s0169-2607(98)00046-7

    authors: Kondoh H,Washiashi T,Sasagaki M,Arisawa J,Nakamura H,Inamura K

    更新日期:1998-08-01 00:00:00