Plenary Talk by G. Volpe at ENO-CANCOA, Cartagena, Colombia, 13 June 2024

DeepTrack 2.1 Logo. (Image from DeepTrack 2.1 Project)
Deep learning for microscopy
Giovanni Volpe
Encuentro Nacional de Óptica y la Conferencia Andina y del Caribe en Óptica y sus Aplicaciones(ENO-CANCOA)
Cartagena, Colombia, 13 June 2024

Video microscopy has a long history of providing insights and breakthroughs for a broad range of disciplines, from physics to biology. Image analysis to extract quantitative information from video microscopy data has traditionally relied on algorithmic approaches, which are often difficult to implement, time consuming, and computationally expensive. Recently, alternative data-driven approaches using deep learning have greatly improved quantitative digital microscopy, potentially offering automatized, accurate, and fast image analysis. However, the combination of deep learning and video microscopy remains underutilized primarily due to the steep learning curve involved in developing custom deep-learning solutions.

To overcome this issue, we have introduced a software, currently at version DeepTrack 2.1, to design, train and validate deep-learning solutions for digital microscopy. We use it to exemplify how deep learning can be employed for a broad range of applications, from particle localization, tracking and characterization to cell counting and classification. Thanks to its user-friendly graphical interface, DeepTrack 2.1 can be easily customized for user-specific applications, and, thanks to its open-source object-oriented programming, it can be easily expanded to add features and functionalities, potentially introducing deep-learning-enhanced video microscopy to a far wider audience.

Deep learning for optical tweezers published in Nanophotonics

Real-time control of optical tweezers with deep learning. (Image by the Authors of the manuscript.)
Deep learning for optical tweezers
Antonio Ciarlo, David Bronte Ciriza, Martin Selin, Onofrio M. Maragò, Antonio Sasso, Giuseppe Pesce, Giovanni Volpe and Mattias Goksör
Nanophotonics, 13(17), 3017-3035 (2024)
doi: 10.1515/nanoph-2024-0013
arXiv: 2401.02321

Optical tweezers exploit light–matter interactions to trap particles ranging from single atoms to micrometer-sized eukaryotic cells. For this reason, optical tweezers are a ubiquitous tool in physics, biology, and nanotechnology. Recently, the use of deep learning has started to enhance optical tweezers by improving their design, calibration, and real-time control as well as the tracking and analysis of the trapped objects, often outperforming classical methods thanks to the higher computational speed and versatility of deep learning. In this perspective, we show how cutting-edge deep learning approaches can remarkably improve optical tweezers, and explore the exciting, new future possibilities enabled by this dynamic synergy. Furthermore, we offer guidelines on integrating deep learning with optical trapping and optical manipulation in a reliable and trustworthy way.

Deep-learning-powered data analysis in plankton ecology published in Limnology and Oceanography Letters

Segmentation of two plankton species using deep learning (N. scintillans in blue, D. tertiolecta in green). (Image by H. Bachimanchi.)
Deep-learning-powered data analysis in plankton ecology
Harshith Bachimanchi, Matthew I. M. Pinder, Chloé Robert, Pierre De Wit, Jonathan Havenhand, Alexandra Kinnby, Daniel Midtvedt, Erik Selander, Giovanni Volpe
Limnology and Oceanography Letters (2024)
doi: 10.1002/lol2.10392
arXiv: 2309.08500

The implementation of deep learning algorithms has brought new perspectives to plankton ecology. Emerging as an alternative approach to established methods, deep learning offers objective schemes to investigate plankton organisms in diverse environments. We provide an overview of deep-learning-based methods including detection and classification of phytoplankton and zooplankton images, foraging and swimming behavior analysis, and finally ecological modeling. Deep learning has the potential to speed up the analysis and reduce the human experimental bias, thus enabling data acquisition at relevant temporal and spatial scales with improved reproducibility. We also discuss shortcomings and show how deep learning architectures have evolved to mitigate imprecise readouts. Finally, we suggest opportunities where deep learning is particularly likely to catalyze plankton research. The examples are accompanied by detailed tutorials and code samples that allow readers to apply the methods described in this review to their own data.

Destructive effect of fluctuations on the performance of a Brownian gyrator published in Soft Matter

Angular velocity in the steady-state. (Excerpt from Fig. 2 of the manuscript.)
Destructive effect of fluctuations on the performance of a Brownian gyrator
Pascal Viot, Aykut Argun, Giovanni Volpe, Alberto Imparato, Lamberto Rondoni, Gleb Oshanin
Soft Matter, 20, 3154-3160 (2024)
arxiv: 2307.05248
doi: 10.1039/D3SM01606D

The Brownian gyrator (BG) is often called a minimal model of a nano-engine performing a rotational motion, judging solely upon the fact that in non-equilibrium conditions its torque, specific angular momentum L and specific angular velocity W have non-zero mean values. For a time-discretised (with time-step δt) model we calculate here the previously unknown probability density functions (PDFs) of L and W. We show that for finite δt, the PDF of L has exponential tails and all moments are therefore well-defined. At the same time, this PDF appears to be effectively broad – the noise-to-signal ratio is generically bigger than unity meaning that L is strongly not self-averaging. Concurrently, the PDF of W exhibits heavy power-law tails and its mean W is the only existing moment. The BG is therefore not an engine in the common sense: it does not exhibit regular rotations on each run and its fluctuations are not only a minor nuisance – on contrary, their effect is completely destructive for the performance. Our theoretical predictions are confirmed by numerical simulations and experimental data. We discuss some plausible improvements of the model which may result in a more systematic rotational motion.

Giovanni Volpe awarded the Göran Gustafsson prize

(Photo by Johan Wingborg.)
Giovanni Volpe was awarded one of Sweden’s most prestigious prizes for physics, the Göran Gustafsson Prize, which is handed out by the Göran Gustafsson Foundation with the help of the Royal Swedish Academy of Sciences. Giovanni receives the physics prize “for boundary breaking research focusing on microscopic particles with active functions”. The prize sum is 6.3 million SEK.

More details here:
Press release of Gothenburg University: Giovanni Volpe receives prestigious Göran Gustafsson prize
Press release of Kungl. Vetenskapsakademien: 33 miljoner till forskning om bland annat TBE och smarta mikropartiklar

Dual-Angle Interferometric Scattering Microscopy for Optical Multiparametric Particle Characterization published in Nano Letters

Conceptual schematic of dual-angle interferometric scattering microscopy (DAISY). (Image by the Authors of the manuscript.)
Dual-Angle Interferometric Scattering Microscopy for Optical Multiparametric Particle Characterization
Erik Olsén, Berenice García Rodríguez, Fredrik Skärberg, Petteri Parkkila, Giovanni Volpe, Fredrik Höök, and Daniel Sundås Midtvedt
Nano Letters, 24(6), 1874-1881 (2024)
doi: 10.1021/acs.nanolett.3c03539
arXiv: 2309.07572

Traditional single-nanoparticle sizing using optical microscopy techniques assesses size via the diffusion constant, which requires suspended particles to be in a medium of known viscosity. However, these assumptions are typically not fulfilled in complex natural sample environments. Here, we introduce dual-angle interferometric scattering microscopy (DAISY), enabling optical quantification of both size and polarizability of individual nanoparticles (radius <170 nm) without requiring a priori information regarding the surrounding media or super-resolution imaging. DAISY achieves this by combining the information contained in concurrently measured forward and backward scattering images through twilight off-axis holography and interferometric scattering (iSCAT). Going beyond particle size and polarizability, single-particle morphology can be deduced from the fact that the hydrodynamic radius relates to the outer particle radius, while the scattering-based size estimate depends on the internal mass distribution of the particles. We demonstrate this by differentiating biomolecular fractal aggregates from spherical particles in fetal bovine serum at the single-particle level.

Colloquium by G. Volpe at the Mini-Symposium with Giovanni Volpe and Pawel Sikorski, Lund, 11 January 2024

(Image by A. Argun)
Deep Learning for Imaging and Microscopy
Giovanni Volpe
Mini-Symposium with Giovanni Volpe and Pawel Sikorski, Lund, Sweden, 11 January 2024
Date: 11 January 2024
Time: 15:15

Video microscopy has a long history of providing insights and breakthroughs for a broad range of disciplines, from physics to biology. Image analysis to extract quantitative information from video microscopy data has traditionally relied on algorithmic approaches, which are often difficult to implement, time consuming, and computationally expensive. Recently, alternative data-driven approaches using deep learning have greatly improved quantitative digital microscopy, potentially offering automatized, accurate, and fast image analysis. However, the combination of deep learning and video microscopy remains underutilized primarily due to the steep learning curve involved in developing custom deep-learning solutions. To overcome this issue, we have introduced a software, DeepTrack 2.1, to design, train and validate deep-learning solutions for digital microscopy.

Symposium on AI, Neuroscience, and Aging featured on ANSA.it

The Symposium on AI, Neuroscience, and Aging has been featured on ANSA.it news, in an article with title: Simposio italo-svedese a Stoccolma sull’IA e la neuroscienza (Italian).

ANSA (an acronym standing for Agenzia Nazionale Stampa Associata) is the leading news agency in Italy and one of the top ranking in the world.

Optimal calibration of optical tweezers with arbitrary integration time and sampling frequencies – A general framework published in Biomedical Optics Express

Different sampling methods for the trajectory of a particle. (Adapted from the manuscript.)
Optimal calibration of optical tweezers with arbitrary integration time and sampling frequencies — A general framework
Laura Pérez-García, Martin Selin, Antonio Ciarlo, Alessandro Magazzù, Giuseppe Pesce, Antonio Sasso, Giovanni Volpe, Isaac Pérez Castillo, Alejandro V. Arzola
Biomedical Optics Express, 14, 6442-6469 (2023)
doi: 10.1364/BOE.495468
arXiv: 2305.07245

Optical tweezers (OT) have become an essential technique in several fields of physics, chemistry, and biology as precise micromanipulation tools and microscopic force transducers. Quantitative measurements require the accurate calibration of the trap stiffness of the optical trap and the diffusion constant of the optically trapped particle. This is typically done by statistical estimators constructed from the position signal of the particle, which is recorded by a digital camera or a quadrant photodiode. The finite integration time and sampling frequency of the detector need to be properly taken into account. Here, we present a general approach based on the joint probability density function of the sampled trajectory that corrects exactly the biases due to the detector’s finite integration time and limited sampling frequency, providing theoretical formulas for the most widely employed calibration methods: equipartition, mean squared displacement, autocorrelation, power spectral density, and force reconstruction via maximum-likelihood-estimator analysis (FORMA). Our results, tested with experiments and Monte Carlo simulations, will permit users of OT to confidently estimate the trap stiffness and diffusion constant, extending their use to a broader set of experimental conditions.

Talk by G. Volpe at the Symposium on AI, Neuroscience, and Aging, Stockholm, 27 November 2023

(Image by A. Argun)
Deep Learning for Imaging and Microscopy
Giovanni Volpe
Symposium on AI, Neuroscience, and Aging, Stockholm, Sweden, 27 November 2023
Date: 27 November 2023
Time: 15:55

Video microscopy has a long history of providing insights and breakthroughs for a broad range of disciplines, from physics to biology. Image analysis to extract quantitative information from video microscopy data has traditionally relied on algorithmic approaches, which are often difficult to implement, time consuming, and computationally expensive. Recently, alternative data-driven approaches using deep learning have greatly improved quantitative digital microscopy, potentially offering automatized, accurate, and fast image analysis. However, the combination of deep learning and video microscopy remains underutilized primarily due to the steep learning curve involved in developing custom deep-learning solutions. To overcome this issue, we have introduced a software, DeepTrack 2.1, to design, train and validate deep-learning solutions for digital microscopy. We use it to exemplify how deep learning can be employed for a broad range of applications, from particle localization, tracking and characterization to cell counting and classification. Thanks to its user- friendly graphical interface, DeepTrack 2.1 can be easily customized for user-specific applications, and, thanks to its open-source object-oriented programming, it can be easily expanded to add features and functionalities, potentially introducing deep-learning-enhanced video microscopy to a far wider audience.