Search

The International Union of Food Scientists (IUFoST) World Congress, a massive global event bringing together over 3,000 delegates, over 80 scientific sessions, over 1000 scientific papers from all over the [...]

This paper provides synthesis methods for large-scale semantic image segmentation datasets of agricultural scenes with the objective to bridge the gap between state-of-the art computer vision performance and that of computer vision in the agricultural robotics domain. We propose a novel methodology to generate renders of random meshes of plants based on empirical measurements, including the automated generation per-pixel class and depth labels for multiple plant parts. A running example is given of Capsicum annuum (sweet or bell pepper) in a high-tech greenhouse. A synthetic dataset of 10,500 images was rendered through Blender, using scenes with 42 procedurally generated plant models with randomised plant parameters. These parameters were based on 21 empirically measured plant properties at 115 positions on 15 plant stems. Fruit models were obtained by 3D scanning and plant part textures were gathered photographically. As reference dataset for modelling and evaluate segmentation performance, 750 empirical images of 50 plants were collected in a greenhouse from multiple angles and distances using image acquisition hardware of a sweet pepper harvest robot prototype. We hypothesised high similarity between synthetic images and empirical images, which we showed by analysing and comparing both sets qualitatively and quantitatively. The sets and models are publicly released with the intention to allow performance comparisons between agricultural computer vision methods, to obtain feedback for modelling improvements and to gain further validations on usability of synthetic bootstrapping and empirical fine-tuning. Finally, we provide a brief perspective on our hypothesis that related synthetic dataset bootstrapping and empirical fine-tuning can be used for improved learning. [ABSTRACT FROM AUTHOR]

Background and aim: Neonatal brain monitoring is increasingly used due to reports of brain injury perioperatively. Little is known about the effect of sedatives (midazolam) and anesthetics (sevoflurane) on cerebral oxygenation (rScO 2 ) and cerebral activity. This study aims to determine these effects in the perioperative period. Methods: This is an observational, prospective study in two tertiary pediatric surgical centers. All neonates with a congenital diaphragmatic hernia received perioperative cerebral oxygenation and activity measurements. Patients were stratified based on intraoperatively administrated medication: the sevoflurane group (continuous sevoflurane, bolus fentanyl, bolus rocuronium) and the midazolam group (continuous midazolam, continuous fentanyl, and continuous vecuronium). Results: Intraoperatively, rScO 2 was higher in the sevoflurane compared to the midazolam group (84%, IQR 77-95 vs. 65%, IQR 59-76, p = < 0.001), fractional tissue oxygen extraction was lower (14%, IQR 5-21 vs. 31%, IQR 29-40, p = < 0.001), the duration of hypoxia was shorter (2%, IQR 0.4-9.6 vs. 38.6%, IQR 4.9-70, p = 0.023), and cerebral activity decreased more: slow delta: 2.16 vs. 4.35 μ V 2 ( p = 0.0049), fast delta: 0.73 vs. 1.37 μ V 2 ( p = < 0.001). In the first 30 min of the surgical procedure, a 3-fold increase in fast delta (10.48-31.22 μ V 2 ) and a 5-fold increase in gamma (1.42-7.58 μ V 2 ) were observed in the midazolam group. Conclusion: Sevoflurane-based anesthesia resulted in increased cerebral oxygenation and decreased cerebral activity, suggesting adequate anesthesia. Midazolam-based anesthesia in neonates with a more severe CDH led to alarmingly low rScO 2 values, below hypoxia threshold, and increased values of EEG power during the first 30 min of surgery. This might indicate conscious experience of pain. Integrating population-pharmacokinetic models and multimodal neuromonitoring are needed for personalized pharmacotherapy in these vulnerable patients. Trial Registration: https://www.trialregister.nl/trial/6972, identifier: NL6972., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2021 Costerus, Hendrikx, IJsselmuiden, Zahn, Perez-Ortiz, Van Huffel, Flint, Caicedo, Wijnen, Wessel, de Graaff, Tibboel and Naulaers.)

Background: Plant parasitic weeds belonging to the genus Striga are a major threat for food production in Sub-Saharan Africa and Southeast Asia. The parasite's life cycle starts with the induction of seed germination by host plant-derived signals, followed by parasite attachment, infection, outgrowth, flowering, reproduction, seed set and dispersal. Given the small seed size of the parasite (< 200 μm), quantification of the impact of new control measures that interfere with seed germination relies on manual, labour-intensive counting of seed batches under the microscope. Hence, there is a need for high-throughput assays that allow for large-scale screening of compounds or microorganisms that adversely affect Striga seed germination., Results: Here, we introduce DiSCount ( Di gital S triga Count er): a computer vision tool for automated quantification of total and germinated Striga seed numbers in standard glass fibre filter assays. We developed the software using a machine learning approach trained with a dataset of 98 manually annotated images. Then, we validated and tested the model against a total dataset of 188 manually counted images. The results showed that DiSCount has an average error of 3.38 percentage points per image compared to the manually counted dataset. Most importantly, DiSCount achieves a 100 to 3000-fold speed increase in image analysis when compared to manual analysis, with an inference time of approximately 3 s per image on a single CPU and 0.1 s on a GPU., Conclusions: DiSCount is accurate and efficient in quantifying total and germinated Striga seeds in a standardized germination assay. This automated computer vision tool enables for high-throughput, large-scale screening of chemical compound libraries and biological control agents of this devastating parasitic weed. The complete software and manual are hosted at https://gitlab.com/lodewijk-track32/discount&#95;paper and the archived version is available at Zenodo with the DOI 10.5281/zenodo.3627138. The dataset used for testing is available at Zenodo with the DOI 10.5281/zenodo.3403956., Competing Interests: Competing interestsLV and JJ are part of Track32, a company specialised in developing computer vision and artificial intelligence applications., (© The Author(s) 2020.)