IGLOO: Machine Vision System for Determination of Solubilization Index in Phosphate-Solubilizing Bacteria

Microorganisms, Journal Year: 2025, Volume and Issue: 13(4), P. 860 - 860

Published: April 9, 2025

Phosphorus is an important macronutrient for plant development, but its bioavailability in soil often limited. Phosphate-solubilizing microorganisms play a vital role phosphorus biogeochemistry, offering sustainable alternative to chemical fertilizers, which pose environmental risks. Manual measurements quantifying phosphate solubilization capacity are laborious, subjective, and time-consuming, so there need develop more efficient objective approaches. This study aimed validate machine vision system called IGLOO automate optimize the determination of relative efficiency phosphate-solubilizing bacteria. was developed using YOLOv8 conjunction with creating labeling dataset images bacterial colonies grown vitro strains Enterobacter R11 FCRK4. The model trained different number epochs. IGLOO’s performance evaluated by comparing segmentation accuracy accepted metrics domain contrasting estimates experts’ manual measurements. achieved greater than 90% colony halo detection, error less 6% compared measurements, demonstrating reliability minimizing observer variability. Finally, represents significant advance quantitative evaluation because it reduces analysis time provides reproducible results agricultural studies.

Language: Английский

Preliminary Study on Sensor-Based Detection of an Adherent Cell’s Pre-Detachment Moment in a MPWM Microfluidic Extraction System

Sensors, Journal Year: 2025, Volume and Issue: 25(9), P. 2726 - 2726

Published: April 25, 2025

The extraction of adherent cells, such as B16 murine melanoma from Petri dish cultures is critical in biomedical applications, including cell reprogramming, transplantation, and regenerative medicine. Traditional detachment methods-enzymatic, mechanical, or chemical-often compromise viability by altering membrane integrity disrupting adhesion proteins. To address these challenges, this study investigated sensor-based detection the pre-detachment phase a MPWM (Microfluidic Pulse Width Modulation) system. Our approach integrates micromechatronic system with microfluidic suction circuit, real-time CCD imaging, computational analysis to detect characterize moment before full extraction. A precisely controlled hydrodynamic force field progressively disrupts multiple stages, reducing mechanical stress preserving integrity. Real-time video enables continuous monitoring positional dynamics oscillatory responses. Image processing deep learning algorithms determine object center coordinates, allowing dynamically adjust parameters. This optimizes while minimizing liquid absorption reflux volume, ensuring efficient By combining microfluidics, sensor detection, AI-driven image processing, established non-invasive method for optimizing detachment. These findings have significant implications single-cell research, medicine, high-throughput biotechnology, maximal minimal perturbation.

Language: Английский

Deep Learning-Based In Situ Micrograph Synthesis and Augmentation for Crystallization Process Image Analysis

Mathematics, Journal Year: 2024, Volume and Issue: 12(22), P. 3448 - 3448

Published: Nov. 5, 2024

Deep learning-based in situ imaging and analysis for crystallization process are essential optimizing product qualities, reducing experimental costs through real-time monitoring, controlling the process. However, large high-quality annotated datasets required to train accurate models, which time consuming. Therefore, we proposed a novel methodology that applied image synthesis neural networks generate virtual information-rich images, enabling efficient rapid dataset expansion while simultaneously annotation costs. Experiments were conducted on L-alanine obtain images validate workflow. The method, aided by interpolation augmentation data warping enhance richness, utilized only 25% of training annotations, consistently segmenting comparable those models utilizing 100% achieving an average precision nearly 98%. Additionally, based Kullback–Leibler divergence, method demonstrated excellent performance extracting information regarding aspect ratios crystal size distributions during Moreover, its ability leverage expert labels with four-fold enhanced efficiency holds great potential advancing various applications processes.

Language: Английский