Deep Learning: Tackling the challenges of bioimage analysis
Deep learning has shown promising results in a wide range of imaging problems in recent years (LeCun et al., 2015), and has the potential to help researchers by automating the analysis of various kinds of biological images (Ronneberger et al., 2015). However, many deep-learning methods require a large amount of 'training data' in order to produce useful results, and this is often not available for bioimage analysis. There is, therefore, a need for deep-learning methods that can make the most from a limited amount of training data. Now, in eLife, Robert Blum (University Hospital Würzburg), Christoph Flath (University of Würzburg) and colleagues – including Dennis Segebarth and Matthias Griebel as joint first authors – provide guidance on how to do this in bioimage analysis (Segebarth et al., 2020).
A common approach to applying deep learning to image analysis involves 'convolutional neural networks': these networks take an input image (such as a microscopy image) and perform many mathematical operations on it to produce an output image (such as a corresponding image with interesting features annotated). A convolutional neural network is characterized by a set of 'learnable parameters', which have to be set to the correct values for the network to perform a given task. The act of finding the correct values for these parameters is called 'training', and several different training techniques are used in practice.
In supervised learning, training is performed using a set of input images and target output images, and the learnable parameters are iteratively adjusted until the output images produced by the network match the target images. It is important to note that supervised learning involves a large amount of randomness, and that training multiple networks using the same data will result in different networks that produce (slightly) different output images.
In bioimage analysis, a common task is to annotate certain structures in images produced by techniques such as microscopy, cryo-EM or X-ray tomography (Meijering et al., 2016). However, the complicated nature of biological images means that this annotation often has to be done by a human expert, which is time-consuming, labour-intensive and subjective (Figure 1A). Supervised deep learning could provide a way to automate the annotation process, reducing the burden on human experts and enabling analysis of a significantly larger set of images. However, annotating the input images needed to train the network also requires a significant amount of time and effort from a human expert.

Different ways to train a convolutional neural network.
Segebarth et al. compare three techniques for training convolutional neural networks to analyze bioimages. (A) In the standard approach a single human expert annotates images for training a single network. (B) In a second approach multiple human experts annotate the same images, and consensus images are used for training: this improves the objectivity of the trained network. (C) In a third approach, a technique called model ensembling is added to the second approach, meaning that multiple networks are trained with the same consensus images: this improves the reliability of the results.
Several properties of a trained network are important for real-world applications. The first is its objectivity, referring to a lack of influences from the subjective nature of human annotations. The second is its reliability, meaning that the trained network should consistently annotate similar features in the same way. The third is its validity, referring to the truthfulness of the network output (that is, did we annotate what we intended to?). In practice, it is often the case that objectivity, reliability and validity are difficult to achieve when the amount of training data is limited.
Various approaches to improve the objectivity, reliability and validity of convolutional neural networks have been proposed. Some involve adapting the structure of the network themselves by, for example, reducing the number of learnable parameters (Pelt and Sethian, 2018), and some involve adapting the training method by, for example, randomly ignoring parts of the network during training (Srivastava et al., 2014). A different approach is to focus on the training data used in supervised learning. Given a network structure and a training method, how can the training data set be optimized to improve objectivity, reliability and validity? In other words, given the time-consuming, labour-intensive and subjective nature of manual annotation, how can a limited period of time from human experts be best utilized to produce a training data set? These questions are currently the subject of active research.
Segebarth et al. investigate two techniques for improving the objectivity, reliability and validity of trained convolutional neural networks in bioimage analysis. First, they investigate the use of multiple human experts to annotate the same set of training images (Figure 1B). The different annotations of each input image are then combined to create a consensus target output image. Since each human expert has their own intended and unintended biases, networks that are trained with data from a single human expert might include the biases of the expert. Using consensus images from multiple experts during training can improve the objectivity of the resulting networks by removing these biases from the training data.
The second technique is to train multiple convolutional neural networks using the same training data set, and then combine the results when the networks are used to analyse new images (Figure 1C). This technique, called model ensembling, has already proven successful in a wide range of applications (Krizhevsky et al., 2017). Model ensembling is based on the randomness involved in training described above: because of this randomness, each trained network will be implicitly biased in their results. By combining the output of multiple networks, these biases are effectively removed, resulting in more reliable results.
A key contribution of Segebarth et al. was to perform extensive experiments on real images and show that the use of consensus images and model ensembles does indeed improve objectivity, reliability and validity. This provides a recipe for optimizing the generation of training data and for making efficient use of the available data, although this recipe still requires a significant amount of human expert time since each image has to be annotated by multiple experts. The results could also help researchers trying to understand how biases affect trained networks, which could lead to improved network structures and training approaches (Müller et al., 2019). And although many questions and challenges remain, the work of Segebarth et al. represents an important step forward in the effort to make the use of deep learning in bioimage analysis feasible.
References
-
ImageNet classification with deep convolutional neural networksCommunications of the ACM 60:84–90.https://doi.org/10.1145/3065386
-
Imagining the future of bioimage analysisNature Biotechnology 34:1250–1255.https://doi.org/10.1038/nbt.3722
-
ConferenceWhen does label smoothing help?Advances in Neural Information Processing Systems. pp. 4694–4703.
-
BookU-net: Convolutional networks for biomedical image segmentationIn: Navab N, Hornegger J, wells W, Frango A, editors. International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer. pp. 234–241.https://doi.org/10.1007/978-3-319-24574-4_28
-
Dropout: a simple way to prevent neural networks from overfittingJournal of Machine Learning Research 15:1929–1958.
Article and author information
Author details
Publication history
- Version of Record published: December 2, 2020 (version 1)
Copyright
© 2020, Pelt
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 1,116
- Page views
-
- 107
- Downloads
-
- 2
- Citations
Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.
Download links
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Further reading
-
- Computational and Systems Biology
- Physics of Living Systems
Diabetes is caused by the inability of electrically coupled, functionally heterogeneous -cells within the pancreatic islet to provide adequate insulin secretion. Functional networks have been used to represent synchronized oscillatory [Ca2+] dynamics and to study -cell subpopulations, which play an important role in driving islet function. The mechanism by which highly synchronized -cell subpopulations drive islet function is unclear. We used experimental and computational techniques to investigate the relationship between functional networks, structural (gap-junction) networks, and intrinsic -cell dynamics in slow and fast oscillating islets. Highly synchronized subpopulations in the functional network were differentiated by intrinsic dynamics, including metabolic activity and KATP channel conductance, more than structural coupling. Consistent with this, intrinsic dynamics were more predictive of high synchronization in the islet functional network as compared to high levels of structural coupling. Finally, dysfunction of gap junctions, which can occur in diabetes, caused decreases in the efficiency and clustering of the functional network. These results indicate that intrinsic dynamics rather than structure drive connections in the functional network and highly synchronized subpopulations, but gap junctions are still essential for overall network efficiency. These findings deepen our interpretation of functional networks and the formation of functional sub-populations in dynamic tissues such as the islet.
-
- Computational and Systems Biology
- Immunology and Inflammation
T cells are required to clear infection, and T cell motion plays a role in how quickly a T cell finds its target, from initial naive T cell activation by a dendritic cell to interaction with target cells in infected tissue. To better understand how different tissue environments affect T cell motility, we compared multiple features of T cell motion including speed, persistence, turning angle, directionality, and confinement of T cells moving in multiple murine tissues using microscopy. We quantitatively analyzed naive T cell motility within the lymph node and compared motility parameters with activated CD8 T cells moving within the villi of small intestine and lung under different activation conditions. Our motility analysis found that while the speeds and the overall displacement of T cells vary within all tissues analyzed, T cells in all tissues tended to persist at the same speed. Interestingly, we found that T cells in the lung show a marked population of T cells turning at close to 180o, while T cells in lymph nodes and villi do not exhibit this “reversing” movement. T cells in the lung also showed significantly decreased meandering ratios and increased confinement compared to T cells in lymph nodes and villi. These differences in motility patterns led to a decrease in the total volume scanned by T cells in lung compared to T cells in lymph node and villi. These results suggest that the tissue environment in which T cells move can impact the type of motility and ultimately, the efficiency of T cell search for target cells within specialized tissues such as the lung.