Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

Cardiology Cardiology

A recurrent neural network using historical data to predict time series indoor PM2.5 concentrations for residential buildings.

In Indoor air

Due to the severe outdoor PM2.5 pollution in China, many people have installed air-cleaning systems in homes. To make the systems run automatically and intelligently, we developed a recurrent neural network (RNN) that uses historical data to predict the future indoor PM2.5 concentration. The RNN architecture includes an autoencoder and a recurrent part. We used data measured in an apartment over the course of an entire year to train and test the RNN. The data include indoor/outdoor PM2.5 concentration, environmental parameters and time of day. By comparing three different input strategies, we found that a strategy employing historical PM2.5 and time of day as inputs performed best. With this strategy, the model can be applied to predict the relatively stable trend of indoor PM2.5 concentration in advance. When the input length is 2 h and the prediction horizon is 30 min, the median prediction error is 8.3 µg/m3 for the whole test set. For times with indoor PM2.5 concentrations between (20,50] µg/m3 and (50,100] µg/m3 , the median prediction error is 8.3 and 9.2 µg/m3 , respectively. The low prediction error between the ground-truth and predicted values shows that the RNN can predict indoor PM2.5 concentrations with satisfactory performance.

Dai Xilei, Liu Junjie, Li Yongle


artificial intelligence, deep learning, indoor PM2.5, outdoor parameters, recurrent neural network, time series model

General General

Quality Control Of Fresh Strawberries By A Random Forest Model.

In Journal of the science of food and agriculture

BACKGROUND : Strawberry quality is one of the most important factors that guarantees consistent commercialization of the fruit and ensures the consumer's satisfaction. This work makes innovative use of Random Forest (RF) to predict sensory measures of strawberries and to classify them in "satisfied" or "not satisfied" and "would pay more" or "wouldn't pay more" using physical and physical-chemical variables. The RF-based model predicts the acceptance, expectation, ideal of sweetness, ideal of acidity, and the ideal of succulence based on the physical and physical-chemical data, which are used as input for the RF-based classification model.

RESULTS : The RF achieved a coefficient of determination R2 > 0.72 and a root-mean-squared error (RMSE) smaller than 0.17 for the prediction task, which indicates that one can estimate the sensory measures of strawberries using physical and physical-chemical data. Furthermore, the RF was able to correctly classify 87.95% of the strawberry samples in the classes "satisfied" and "not satisfied" and 78.99% in the classes "would pay more" or "wouldn't pay more". Additionally, a two-step RF model, which employed both physical and physical-chemical data to classify strawberry samples regarding the consumer's response, correctly classified 100% and 90.32% of the samples with respect to the consumer's satisfaction and its willingness to pay more, respectively.

CONCLUSION : The results indicate that the developed models can be used in the quality control of strawberries, supporting the establishment of quality standards that consider the consumer's response. Additionally, the proposed methodology can be extended to control the sensory quality of other fruits. This article is protected by copyright. All rights reserved.

Ribeiro Michele N, Carvalho Iago A, Fonseca Gabriel A, Lago Rafael C, Rocha Lenízy C R, Ferreira Danton D, Vilas Boas Eduardo V B, Pinheiro Ana C M


classification, machine learning, random forests, regression, sensory response, strawberry

General General

Automatic Lumen Border Detection in IVUS Images Using Deep Learning Model and Handcrafted Features.

In Ultrasonic imaging

In the clinical analysis of Intravascular ultrasound (IVUS) images, the lumen size is an important indicator of coronary atherosclerosis, and is also the premise of coronary artery disease diagnosis and interventional treatment. In this study, a fully automatic method based on deep learning model and handcrafted features is presented for the detection of the lumen borders in IVUS images. First, 193 handcrafted features are extracted from the IVUS images. Then hybrid feature vectors are constructed by combining handcrafted features with 64 high-level features extracted from U-Net. In order to obtain the feature subsets with larger contribution, we employ the extended binary cuckoo search for feature selection. Finally, the selected 36-dimensional hybrid feature subset is used to classify the test images using dictionary learning based on kernel sparse coding. The proposed algorithm is tested on the publicly available dataset and evaluated using three indicators. Through ablation experiments, mean value of the experimental results (Jaccard: 0.88, Hausdorff distance: 0.36, Percentage of the area difference: 0.06) prove to be effective improving lumen border detection. Furthermore, compared with the recent methods used on the same dataset, the proposed method shows good performance and high accuracy.

Li Kai, Tong Jijun, Zhu Xinjian, Xia Shudong


deep learning, dictionary learning, feature selection, intravascular ultrasound image, lumen border detection

General General

Unsupervised Machine Learning-Based Clustering of Nanosized Fluorescent Extracellular Vesicles.

In Small (Weinheim an der Bergstrasse, Germany)

Extracellular vesicles (EV) are biological nanoparticles that play an important role in cell-to-cell communication. The phenotypic profile of EV populations is a promising reporter of disease, with direct clinical diagnostic relevance. Yet, robust methods for quantifying the biomarker content of EV have been critically lacking, and require a single-particle approach due to their inherent heterogeneous nature. Here, multicolor single-molecule burst analysis microscopy is used to detect multiple biomarkers present on single EV. The authors classify the recorded signals and apply the machine learning-based t-distributed stochastic neighbor embedding algorithm to cluster the resulting multidimensional data. As a proof of principle, the authors use the method to assess both the purity and the inflammatory status of EV, and compare cell culture and plasma-derived EV isolated via different purification methods. This methodology is then applied to identify intercellular adhesion molecule-1 specific EV subgroups released by inflamed endothelial cells, and to prove that apolipoprotein-a1 is an excellent marker to identify the typical lipoprotein contamination in plasma. This methodology can be widely applied on standard confocal microscopes, thereby allowing both standardized quality assessment of patient plasma EV preparations, and diagnostic profiling of multiple EV biomarkers in health and disease.

Kuypers Sören, Smisdom Nick, Pintelon Isabel, Timmermans Jean-Pierre, Ameloot Marcel, Michiels Luc, Hendrix Jelle, Hosseinkhani Baharak


burst analysis spectroscopy, extracellular vesicles, machine learning, multidimensional phenotyping

General General

Deep learning magnetic resonance spectroscopy fingerprints of brain tumours using quantum mechanically synthesised data.

In NMR in biomedicine ; h5-index 41.0

Metabolic fingerprints are valuable biomarkers for diseases that are associated with metabolic disorders. 1H magnetic resonance spectroscopy (MRS) is a unique noninvasive diagnostic tool that can depict the metabolic fingerprint based solely on the proton signal of different molecules present in the tissue. However, its performance is severely hindered by low SNR, field inhomogeneities and overlapping spectra of metabolites, which affect the quantification of metabolites. Consequently, MRS is rarely included in routine clinical protocols and has not been proven in multi-institutional trials. This work proposes an alternative approach, where instead of quantifying metabolites' concentration, deep learning (DL) is used to model the complex nonlinear relationship between diseases and their spectroscopic metabolic fingerprint (pattern). DL requires large training datasets, acquired (ideally) with the same protocol/scanner, which are very rarely available. To overcome this limitation, a novel method is proposed that can quantum mechanically synthesise MRS data for any scanner/acquisition protocol. The proposed methodology is applied to the challenging clinical problem of differentiating metastasis from glioblastoma brain tumours on data acquired across multiple institutions. DL algorithms were trained on the augmented synthetic spectra and tested on two independent datasets acquired by different scanners, achieving a receiver operating characteristic area under the curve of up to 0.96 and 0.97, respectively.

Dikaios Nikolaos


deep learning, magnetic resonance spectroscopy, metabolic fingerprint, quantum mechanical spectroscopy, synthetic data

General General

A Deep Learning Method for Alerting Emergency Physicians about the Presence of Subphrenic Free Air on Chest Radiographs.

In Journal of clinical medicine

Hollow organ perforation can precipitate a life-threatening emergency due to peritonitis followed by fulminant sepsis and fatal circulatory collapse. Pneumoperitoneum is typically detected as subphrenic free air on frontal chest X-ray images; however, treatment is reliant on accurate interpretation of radiographs in a timely manner. Unfortunately, it is not uncommon to have misdiagnoses made by emergency physicians who have insufficient experience or who are too busy and overloaded by multitasking. It is essential to develop an automated method for reviewing frontal chest X-ray images to alert emergency physicians in a timely manner about the life-threatening condition of hollow organ perforation that mandates an immediate second look. In this study, a deep learning-based approach making use of convolutional neural networks for the detection of subphrenic free air is proposed. A total of 667 chest X-ray images were collected at a local hospital, where 587 images (positive/negative: 267/400) were used for training and 80 images (40/40) for testing. This method achieved 0.875, 0.825, and 0.889 in sensitivity, specificity, and AUC score, respectively. It may provide a sensitive adjunctive screening tool to detect pneumoperitoneum on images read by emergency physicians who have insufficient clinical experience or who are too busy and overloaded by multitasking.

Su Che-Yu, Tsai Tsung-Yu, Tseng Cheng-Yen, Liu Keng-Hao, Lee Chi-Wei


convolutional neural networks, emergency physicians, frontal chest X-ray images, hollow organ perforation, subphrenic free air