1.1 Introduction

Bioimage analysis is often regarded as a technical task that can be solved simply by the development of new and more sophisticated image processing algorithms. This may be true to a large extent, but the complexity encountered in the actual usage of those algorithms during the analysis leads to a number of challenges that leave researchers with a thought that “Bioimage analysis is difficult”.

To provide structure and organization in this complexity, and by that to enable and simplify users’ navigation through it, the Network of European Bioimage Analysts (NEUBIAS) has been systematically looking at the computational tools and algorithmic resources of bioimage analysis with a slightly higher resolution, identifying components, collections, and workflows (Miura et al. , 2020). Each of these terms define different types of computational tools. Component is an implementation of a certain image processing or analysis algorithm; Collection is a software package, or a library, that includes many valuable (possibly independent) components, and is offered as a collection of downloadable files, ready to be used; Workflow is created by assembling the components (e.g., from one or more collections) into a sequence of image processing and analysis steps, to solve a certain biological question. Workflows typically take raw image data as input and aim at delivering parameters of biological systems and/or visualization of the system analysis results as an output.

For creating a workflow, knowledge of the characteristics of various components and their behavior against image data is required. At the same time, one needs to know some standard methods for assembling components into workflows. Furthermore, the ability of a user to use programming language becomes mandatory, as it dramatically enhances the range of components one can select from, and increases the efficiency of automated analysis. Moreover, presenting a workflow in the form of a computer program can be regarded as a highly recommended scientific practice for method reproducibility. Therefore, the training in bioimage analysis should ideally include the three main elements: component-related literacy, programming language fluency, and workflow design.

In the previous textbook (Miura and Sladoje , 2020) prepared by, and for, the NEUBIAS community, as well as in the earlier BIAS textbook (Miura , 2016b), we focused primarily on introducing the main principles of workflow design, and how to implement workflows using scripting languages such as ImageJ Macro, MATLAB, and R. We selected this particular approach with an aim to reduce the imbalance between the vast amount of already existing literature and textbooks focused on image processing and analysis algorithms (i.e. components), in comparison with scarce resources for learning how to design and implement bioimage analysis workflows. Contributing authors were asked to provide a holistic view of a bioimage analysis task, starting with introducing the biological background relevant for their chapter, and describing the biological research question that they want to address by a fully-coded and reproducible image analysis workflow. In addition, they were expected to provide a detailed explanation of the code, and – finally – interpretation of the results of the performed analysis, in terms of the biological question in focus. These contributions narrowed the gap, at least to some extent as we believe, between the ever growing number, excellence, and complexity of image analysis components on one side, and the biological questions to be addressed by them, on the other side. The textbook was warmly endorsed by the community of life scientists, as well as bioimage analysts, who have been using it as a valuable resource.

The Network of European Bioimage Analysts has been continuously growing, in terms of size and competence. Via education and communication, supported by numerous training sessions, dedicated conferences, research collaborations, discussion forums and several other activities, our members have learned many of the techniques used in the highly multidisciplinary field of bioimage analysis. To best respond to their needs, we have decided to widen the scope of this new Bioimage Analysis textbook in two ways.

Firstly, we have included several chapters devoted to components, in response to the increasing demand to use the cutting-edge algorithms and follow the most recent trends in the field of bioimage analysis. In our opinion, this demand is a direct consequence of the narrowed gap in interdisciplinary competences and communication between life scientists (biologists) and computer scientists, and increased competences of bioimage analysts who have become more skilled in bridging those two fields. We appreciate this as a valuable outcome of various efforts made by the NEUBIAS community: Increased interest and utilization of high-end components in life sciences result from the presence of the new experts. Note that, however, the authors of these “component” chapters preserved the main flavour of our textbooks – the biological context, exemplified by use-cases of the presented components, possibly within workflows.

Secondly, the increase in the number of bioimage analysts, but also in the level of their skills and competences, has motivated us to include another novel form of contributions: “workflow deconstruction” chapters. This pedagogical approach in bioimage analysis training has been proposed by Jean-Yves Tinevez and Kota Miura, inspired by the “Deconstruction” concept introduced by postmodern philosopher Jaques Derrida, and developed during the NEUBIAS Training Schools for Bioimage Analysts in several editions between years 2016 and 2020. The aim has been to maximize the learning experience in workflow design by learning to generalize the knowledge and techniques gained from a small sample of well selected examples of workflows, deconstructed and discussed in detail with respect to how components are assembled and critically evaluated within the design. This approach was suitable for the trainees proficient in computer programming and experienced in usage of a variety of components; these were primarily professional bioimage analysts. We hope that “workflow deconstruction” chapters included in this book provide insight in the essence of workflow design, and also ignite readers’ creativity in suggesting their own novel bioimage analysis workflows.

As a result, this Volume 2 collection includes seven chapters. The book starts with discussion on “Batch Processing Methods in ImageJ” (Chap. 1), and presentation of tools available in “Python: Data handling, analysis and plotting” (Chap. 2), both aiming to increase the fluency in programming languages, “tidy” data handling, and environments widely used in bioimage analysis. The subsequent chapters are focused on components: “Building a Bioimage Analysis Workflow using Deep Learning” (Chap. 3) and “GPU-accelerating ImageJ Macro image processing workflows using CLIJ” (Chap. 4); both describe ways to include cutting-edge components into a variety of workflows, responding to clear demands from the bioimage analysis community. We continue, and conclude, with three chapters devoted to workflow deconstruction, putting in focus three different biological problems, and suggesting and analysing their original suggested solutions: “SurfCut macro deconstruction” (Chap. 5), “i.2.i. with the (fruit) fly: Quantifying position effect variegation in Drosophila melanogaster” (Chap. 6), and “A Matlab pipeline for spatiotemporal quantification of monolayer cell migration” (Chap. 7). These chapters require certain literacy in programming, but offer numerous valuable tips out of which many are generally applicable. In particular, Chap. 5 aims to provide an introduction to the concept and practice of “workflow deconstruction”, demonstrating the process in detail. Chapters 6 and 7 follow with examples of very successful original designs and utilization of image analysis workflows to perform detailed and unique analysis of extracted biological parameters.

Chapters 1, 3, 4, 5, and 6 require some basic knowledge of ImageJ macro language. If lacking it, the readers are referred to “ImageJ Macro Language” (Miura , 2016a). Chapters 2 and 3 assume basic knowledge of Python programming. Chapter 7 requires basic knowledge of MATLAB programming. There are numerous available resources to support readers to meet these requirements; in particular, we mention “Introduction to MATLAB” (Monzel and Möhl , 2016) and “Introduction to MATLAB” (Nørrelykke , 2020), the former being general and basic, and the latter slightly more advanced.

This textbook is the 2nd bioimage analysis textbook published as an output of the common efforts of NEUBIAS, funded under COST Action CA15124. We would like to thank the project workgroup (WG) leaders: Sebastian Munck, Arne Seitz, and Florian Levet (WG1 “Strategy”); Paula Sampaio and Irene Fondón (WG2 “Outreach”); Gaby Martins and Fabrice Cordeliéres (WG3 “Training); Perrine Paul-Gilloteaux and Chong Zhang (WG4 “Webtool biii.eu”); Sébastien Tosi, Graeme Ball and Raphaël Marée (WG5 “Benchmarking and Sample Datasets”); Julia Fernandez-Rodriguez and Clara Prats Gavalda (WG7 “Short-Term Scientific Missions and Career Path”); and Julien Colombelli (the Action Chair). Their efforts to create a synergistic effect of the diverse workgroup activities towards the establishment of “Bioimage Analysts” is the strong backbone that has led to the successful realization of this book as a result of WG6 “Open Publication” (led by Editors). We are very much grateful to the reviewers of each chapter: Jan Eglinger, Uwe Schmidt, Martin Weigert, Sebastién Tosi, Dominic Waithe, Jonas Øgaard, Mafalda Sousa, and Simon F. Nørrelykke. Their critical comments largely improved the presented content. We are particularly grateful to the authors of each chapter: Anna Klemm, Kota Miura, Arianne Bercowsky Rama, Estibaliz Gomez-de-Mariscal, Daniel Franco-Barranco, Arrate Muñoz-Barrutia, Ignacio Arganda-Carreras, Daniela Vorkel, Robert Haase, Bertrand Cinquin, Joyce Y. Kao, Mark L. Siegal, Marion Louveaux, Stephane Verger, Yishaia Zabary, and Assaf Zaritsky; for their selfless commitment to meet the demanding requirements of the publication format that we have chosen. The publication of this book was enabled by the financial support from the COST Association (funded through EU framework Horizon2020), through the granted project “A New Network of European Bioimage Analysts (NEUBIAS, COST Action CA15124)”. Finally, we wish to thank all members of NEUBIAS who, with their enthusiasm and commitment to the network’s diverse activities, have contributed to keep the momentum of the initiative constantly high, a vital element to enable it to reach its objectives, including the publication of this book.