Convolutional neural networks in the qualitative improvement of sweet potato roots.

The objective was to verify whether convolutional neural networks can help sweet potato phenotyping for qualitative traits. We evaluated 16 families of sweet potato half-sibs in a randomized block design with four replications. We obtained the images at the plant level and used the ExpImage package of the R software to reduce the resolution and individualize one root per image. We grouped them according to their classifications regarding shape, peel color, and damage caused by insects. 600 roots of each class were destined for training the networks, while the rest was used to verify the quality of the fit. We used the python language on the Google Colab platform and the Keras library, considering the VGG-16, Inception-v3, ResNet-50, InceptionResNetV2, and EfficientNetB3 architectures. The InceptionResNetV2 architecture stood out with high accuracy in classifying individuals according to shape, insect damage, and peel color. Image analysis associated with deep learning may help develop applications used by rural producers and improve sweet potatoes, reducing subjectivity, labor, time, and financial resources in phenotyping.

Utility of echocardiographic right ventricular subcostal strain in critical care.

AIMS: Right ventricular (RV) strain is a known predictor of outcomes in various heart and lung pathologies but has been considered too technically challenging for routine use in critical care. We examined whether RV strain acquired from the subcostal view, frequently more accessible in the critically ill, is an alternative to conventionally derived RV strain in intensive care. METHODS AND RESULTS: RV strain data were acquired from apical and subcostal views on transthoracic echocardiography (TTE) in 94 patients (35% female), mean age 50.5 ± 15.2 years, venovenous extracorporeal membrane oxygenation (VVECMO) (44%). RV strain values from the apical (mean ± standard deviation; -20.4 ± 6.7) and subcostal views (-21.1 ± 7) were highly correlated (Pearson's r -0.89, P < 0.001). RV subcostal strain correlated moderately well with other echocardiography parameters including tricuspid annular plane systolic excursion (r -0.44, P < 0.001), RV systolic velocity (rho = -0.51, P < 0.001), fractional area change (r -0.66, P < 0.01), and RV outflow tract velocity time integral (r -0.49, P < 0.001). VVECMO was associated with higher RV subcostal strain (non-VVECMO -19.6 ± 6.7 vs. VVECMO -23.2 ± 7, P = 0.01) but not apical RV strain. On univariate analysis, RV subcostal strain was weakly associated with survival at 30 days (R2 = 0.04, P = 0.05, odds ratio =1.08) while apical RV was not (P = 0.16). CONCLUSION: RV subcostal deformation imaging is a reliable surrogate for conventionally derived strain in critical care and may in time prove to be a useful diagnostic marker in this cohort.