Stitched vision transformer for age-related macular degeneration detection using retinal optical coherence tomography images.

Age-related macular degeneration (AMD) is an eye disease that leads to the deterioration of the central vision area of the eye and can gradually result in vision loss in elderly individuals. Early identification of this disease can significantly impact patient treatment outcomes. Furthermore, given the increasing elderly population globally, the importance of automated methods for rapidly monitoring at-risk individuals and accurately diagnosing AMD is growing daily. One standard method for diagnosing AMD is using optical coherence tomography (OCT) images as a non-invasive imaging technology. In recent years, numerous deep neural networks have been proposed for the classification of OCT images. Utilizing pre-trained neural networks can speed up model deployment in related tasks without compromising accuracy. However, most previous methods overlook the feasibility of leveraging pre-existing trained networks to search for an optimal architecture for AMD staging on a new target dataset. In this study, our objective was to achieve an optimal architecture in the efficiency-accuracy trade-off for classifying retinal OCT images. To this end, we employed pre-trained medical vision transformer (MedViT) models. MedViT combines convolutional and transformer neural networks, explicitly designed for medical image classification. Our approach involved pre-training two distinct MedViT models on a source dataset with labels identical to those in the target dataset. This pre-training was conducted in a supervised manner. Subsequently, we evaluated the performance of the pre-trained MedViT models for classifying retinal OCT images from the target Noor Eye Hospital (NEH) dataset into the normal, drusen, and choroidal neovascularization (CNV) classes in zero-shot settings and through five-fold cross-validation. Then, we proposed a stitching approach to search for an optimal model from two MedViT family models. The proposed stitching method is an efficient architecture search algorithm known as stitchable neural networks. Stitchable neural networks create a candidate model in search space for each pair of stitchable layers by inserting a linear layer between them. A pair of stitchable layers consists of layers, each selected from one input model. While stitchable neural networks had previously been tested on more extensive and general datasets, this study demonstrated that stitching networks could also be helpful in smaller medical datasets. The results of this approach indicate that when pre-trained models were available for OCT images from another dataset, it was possible to achieve a model in 100 epochs with an accuracy of over 94.9% in classifying images from the NEH dataset. The results of this study demonstrate the efficacy of stitchable neural networks as a fine-tuning method for OCT image classification. This approach not only leads to higher accuracy but also considers architecture optimization at a reasonable computational cost.

Automated identification of toxigenic cyanobacterial genera for water quality control purposes.

Cyanobacteria are the dominating microorganisms in aquatic environments, posing significant risks to public health due to toxin production in drinking water reservoirs. Traditional water quality assessments for abundance of the toxigenic genera in water samples are both time-consuming and error-prone, highlighting the urgent need for a fast and accurate automated approach. This study addresses this gap by introducing a novel public dataset, TCB-DS (Toxigenic Cyanobacteria Dataset), comprising 2593 microscopic images of 10 toxigenic cyanobacterial genera and subsequently, an automated system to identify these genera which can be divided into two parts. Initially, a feature extractor Convolutional Neural Network (CNN) model was employed, with MobileNet emerging as the optimal choice after comparing it with various other popular architectures such as MobileNetV2, VGG, etc. Secondly, to perform classification algorithms on the extracted features of the first section, multiple approaches were tested and the experimental results indicate that a Fully Connected Neural Network (FCNN) had the optimal performance with weighted accuracy and f1-score of 94.79% and 94.91%, respectively. The highest macro accuracy and f1-score were 90.17% and 87.64% which were acquired using MobileNetV2 as the feature extractor and FCNN as the classifier. These results demonstrate that the proposed approach can be employed as an automated screening tool for identifying toxigenic Cyanobacteria with practical implications for water quality control replacing the traditional estimation given by the lab operator following microscopic observations. The dataset and code of this paper are publicly available at https://github.com/iman2693/CTCB.

Contribution of machine learning approaches in response to SARS-CoV-2 infection.

PROBLEM: The lately emerged SARS-CoV-2 infection, which has put the whole world in an aberrant demanding situation, has generated an urgent need for developing effective responses through artificial intelligence (AI). AIM: This paper aims to overview the recent applications of machine learning techniques contributing to prevention, diagnosis, monitoring, and treatment of coronavirus disease (SARS-CoV-2). METHODS: A progressive investigation of the recent publications up to November 2020, related to AI approaches towards managing the challenges of COVID-19 infection was made. RESULTS: For patient diagnosis and screening, Convolutional Neural Network (CNN) and Support Vector Machine (SVM) are broadly applied for classification purposes. Moreover, Deep Neural Network (DNN) and homology modeling are the most used SARS-CoV-2 drug repurposing models. CONCLUSION: While the fields of diagnosis of the SARS-CoV-2 infection by medical image processing and its dissemination pattern through machine learning have been sufficiently studied, some areas such as treatment outcome in patients and drug development need to be further investigated using AI approaches.

Potential of blockchain approach on development and security of microbial databases.

Approaches developed based on the blockchain concept can provides a framework for the realization of open science. The traditional centralized way of data collection and curation is a labor-intensive work that is often not updated. The fundamental contribution of developing blockchain format of microbial databases includes: 1. Scavenging the sparse data from different strain database; 2. Tracing a specific thread of access for the purpose of evaluation or even the forensic; 3. Mapping the microbial species diversity; 4. Enrichment of the taxonomic database with the biotechnological applications of the strains and 5. Data sharing with the transparent way of precedent recognition. The plausible applications of constructing microbial databases using blockchain technology is proposed in this paper. Nevertheless, the current challenges and constraints in the development of microbial databases using the blockchain module are discussed in this paper.

Spectrum of deep learning algorithms in drug discovery.

Deep learning (DL) algorithms are a subset of machine learning algorithms with the aim of modeling complex mapping between a set of elements and their classes. In parallel to the advance in revealing the molecular bases of diseases, a notable innovation has been undertaken to apply DL in data/libraries management, reaction optimizations, differentiating uncertainties, molecule constructions, creating metrics from qualitative results, and prediction of structures or interactions. From source identification to lead discovery and medicinal chemistry of the drug candidate, drug delivery, and modification, the challenges can be subjected to artificial intelligence algorithms to aid in the generation and interpretation of data. Discovery and design approach, both demand automation, large data management and data fusion by the advance in high-throughput mode. The application of DL can accelerate the exploration of drug mechanisms, finding novel indications for existing drugs (drug repositioning), drug development, and preclinical and clinical studies. The impact of DL in the workflow of drug discovery, design, and their complementary tools are highlighted in this review. Additionally, the type of DL algorithms used for this purpose, and their pros and cons along with the dominant directions of future research are presented.

Automated identification of Myxobacterial genera using Convolutional Neural Network.

The Myxococcales order consist of eleven families comprising30 genera, and are featured by the formation of the highest level of differential structure aggregations called fruiting bodies. These multicellular structures are essential for their resistance in ecosystems and is used in the primitive identification of these bacteria while their accurate taxonomic position is confirmed by the nucleotide sequence of 16SrRNA gene. Phenotypic classification of these structures is currently performed based on the stereomicroscopic observations that demand personal experience. The detailed phenotypic features of the genera with similar fruiting bodies are not readily distinctive by not particularly experienced researchers. The human examination of the fruiting bodies requires high skill and is error-prone. An image pattern analysis of schematic images of these structures conducted us to the construction of a database, which led to an extractable recognition of the unknown fruiting bodies. In this paper, Convolutional Neural Network (CNN) was considered as a baseline for recognition of fruiting bodies. In addition, to enhance the result the classifier, part of CNN is replaced with other classifiers. By employing the introduced model, all 30 genera of this order could be recognized based on stereomicroscopic images of the fruiting bodies at the genus level that not only does not urge us to amplify and sequence gene but also can be attained without preparation of microscopic slides of the vegetative cells or myxospores. The accuracy of 77.24% in recognition of genera and accuracy of 88.92% in recognition of suborders illustrate the applicability property of the proposed machine learning model.

Age Prediction Based on Brain MRI Image: A Survey.

Human age prediction is an interesting and applicable issue in different fields. It can be based on various criteria such as face image, DNA methylation, chest plate radiographs, knee radiographs, dental images and etc. Most of the age prediction researches have mainly been based on images. Since the image processing and Machine Learning (ML) techniques have grown up, the investigations were led to use them in age prediction problem. The implementations would be used in different fields, especially in medical applications. Brain Age Estimation (BAE) has attracted more attention in recent years and it would be so helpful in early diagnosis of some neurodegenerative diseases such as Alzheimer, Parkinson, Huntington, etc. BAE is performed on Magnetic Resonance Imaging (MRI) images to compute the brain ages. Studies based on brain MRI shows that there is a relation between accelerated aging and accelerated brain atrophy. This refers to the effects of neurodegenerative diseases on brain structure while making the whole of it older. This paper reviews and summarizes the main approaches for age prediction based on brain MRI images including preprocessing methods, useful tools used in different research works and the estimation algorithms. We categorize the BAE methods based on two factors, first the way of processing MRI images, which includes pixel-based, surface-based, or voxel-based methods and second, the generation of ML algorithms that includes traditional or Deep Learning (DL) methods. The modern techniques as DL methods help MRI based age prediction to get results that are more accurate. In recent years, more precise and statistical ML approaches have been utilized with the help of related tools for simplifying computations and getting accurate results. Pros and cons of each research and the challenges in each work are expressed and some guidelines and deliberations for future research are suggested.

An image analysis-aided method for redundancy reduction in differentiation of identical Actinobacterial strains.

AIM: To simplify the recognition of Actinobacteria, at different stages of the growth phase, from a mixed culture to facilitate the isolation of novel strains of these bacteria for drug discovery purposes. MATERIALS & METHODS: A method was developed based on Gabor transform, and machine learning using k-Nearest Neighbors and Naive Bayes classifier, Logitboost, Bagging and Random Forest to automatically categorize the colonies. RESULTS: A signature pattern was inferred by the model, making the differentiation of identical strains possible. Additionally, higher performance, compared with other classification methods was achieved. CONCLUSION: This automated approach can contribute to the acceleration of the drug discovery process while it simultaneously can diminish the loss of budget due to the redundancy occurred by the inexperienced researchers.