Open Source Lab Manual

[Lyle Roblez]

Leveragebest practices for running an open source program office or starting an open source project in your organization. Developed by The Linux Foundation in partnership with the TODO Group, these resources represent the experience of our staff, projects, and members.

Open source program managers must demonstrate the ROI of their efforts. This guide provides an overview of some of the standard ways that organizations evaluate their open source programs, projects, and contributions. Learn what to measure, how to define success, and how to best use this information to advance your open source program objectives, ...

One of the most important responsibilities of an open source program office is ensuring that your organization meets its legal obligations when integrating open source code with proprietary and third-party source code in your commercial products. You need to establish guidelines on how developers can use open source code, and detailed processes to ...

Open source development requires a different approach than many organizations are accustomed to. But it becomes easier if you have a clear plan to follow. Fortunately, so many companies and individuals have already forged a path to success in contributing to significant open source projects. They have tried and true methods for establishing a ...

This Open Source Guide is designed to offer advice about how your enterprise and your development team can plan for the day when you are ready to end or move away from an unneeded open source project. By shutting down the project gracefully or by transitioning it to others who can continue the work, your enterprise can responsibly oversee the life ...

Integrating into open source communities takes time and effort and requires a new approach to product development. Where traditional, proprietary development requires secrecy and a management hierarchy, open source development requires openness and values consensus. Code contributions, not title or position, are what determine influence and ...

The majority of companies that use open source understand its business value and identify its advantages in efficiency, flexibility, interoperability, and speed of innovation. Yet only half of these companies report practicing basic open source management, such as community development, code maintenance, and the like, according to the latest ...

Open source is about more than just code. It's also about the planning thathappens before the code is written, the process of how that code is used byothers, and fostering a welcoming environment where a community can grow.

In the spirit of openness, we are publishing our internal documentation for howwe do open source at Google. We invite you to take a look behind the scenes athow we use, release, and support open source projects and communities.

This is a copy of our internal open source documentation, with a few exceptions.For a number of reasons, we can't share everything, so you might find placeswhere a link is missing or some content had to be removed.

Aside from those few cases, this is the same documentation seen by Googleemployees. As a result, there is some Google lingo throughout, as well asreferences to internal tools and systems. See the glossary for definitions ofsome of the most common ones.

Creating covers how Googlersrelease code that they've written, either in the form of a new standaloneproject or as a patch to an external project. The same process is used forsmall 20% projects and full blown Google projects.

For other companies that are releasing or using open source software, we want toshare the lessons we've learned from many years of experience. By being astransparent as we can about how we do open source, we hope to help others do thesame.

However, many of the things we do are unique to how Google operates and ourengineering culture, so these should not be read as "how-to" guides. To hearfrom more companies deeply involved in open source, we recommend checking outthe TODO Group.

For individuals or project maintainers, if you've ever contributed to Googleopen source projects, or received a patch from a Googler, you've been exposed tohow we do things. We hope that these documents provide useful insight into howwe approach open source and answer questions you may have.

Yes, we really do publish our internal documentation for all the world to see!One of the goals of the site is to expose (where appropriate) our internalprocesses related to open source so that other companies have the opportunity tolearn from them.

We carefully review all content before it's published externally, usingautomated tooling (go/cleanr) to rewrite internal URLs, file paths, and certainproject code names. We also perform a human review of all changes before they'republished. In general, we don't consider go/ links to be sensitive since theyaren't accessible by non-Googlers and typically have generic names.

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Stroke is the leading cause of adult disability worldwide, with up to two-thirds of individuals experiencing long-term disabilities. Large-scale neuroimaging studies have shown promise in identifying robust biomarkers (e.g., measures of brain structure) of long-term stroke recovery following rehabilitation. However, analyzing large rehabilitation-related datasets is problematic due to barriers in accurate stroke lesion segmentation. Manually-traced lesions are currently the gold standard for lesion segmentation on T1-weighted MRIs, but are labor intensive and require anatomical expertise. While algorithms have been developed to automate this process, the results often lack accuracy. Newer algorithms that employ machine-learning techniques are promising, yet these require large training datasets to optimize performance. Here we present ATLAS (Anatomical Tracings of Lesions After Stroke), an open-source dataset of 304 T1-weighted MRIs with manually segmented lesions and metadata. This large, diverse dataset can be used to train and test lesion segmentation algorithms and provides a standardized dataset for comparing the performance of different segmentation methods. We hope ATLAS release 1.1 will be a useful resource to assess and improve the accuracy of current lesion segmentation methods.

Approximately 795,000 people in the United States suffer from a stroke every year, resulting in nearly 133,000 deaths1. In addition, up to 2/3 of stroke survivors experience long-term disabilities that impair their participation in daily activities2,3. Careful clinical decision making is thus critical both at the acute stage, where interventions can spare neural tissue or be used to promote early functional recovery4, and at the subacute/chronic stages, where effective rehabilitation can promote long-term functional recovery. Enormous efforts have been made to predict outcomes and response to treatments at both acute and subacute/chronic stages using brain imaging.

Many stroke neuroimaging studies have utilized semi- or fully-automated lesion segmentation tools for their analyses. Semi-automated segmentation tools employ a combination of automated algorithms, which detect abnormalities in the MR image, and manual corrections or inputs by an expert. Fully-automated algorithms rely completely on the algorithm for the lesion segmentation. While these require little human input or expertise, they still may require significant computational resources and processing time. Many of these fully-automated algorithms employ machine learning techniques that require training and testing on large datasets21, and the performance of the algorithm is highly dependent on the size and diversity of the training dataset. While there have been several exciting initiatives regarding lesion segmentation in acute clinical imaging, discussed below, there are few publically available large training/test datasets of manually segmented stroke lesion masks on research-grade T1-weighted images that could be used for improving such algorithms. Thus, while both semi- and fully-automated lesion segmentation tools have the potential to greatly reduce the time and expertise needed to analyze stroke MRI data22, it is unclear whether they provide the accuracy needed for rigorous stroke lesion-based analyses.

Here, we present ATLAS (Anatomical Tracings of Lesions After Stroke) Release 1.1, an open-source dataset consisting of 304 T1-weighted MRIs with manually segmented diverse lesions and metadata. The goal of ATLAS is to provide the research community with a standardized training and testing dataset for lesion segmentation algorithms on T1-weighted MRIs. We note that this dataset is not representative of the full range of stroke, as this data was acquired through research studies in which individuals with stroke voluntarily participated, and all participants had to be eligible for a research MRI session. However, this dataset may be useful for testing and comparing the performance of different lesion segmentation techniques and identifying key barriers hindering the performance of automated lesion segmentation algorithms. We believe that this diverse set of manually segmented lesions will serve as a valuable resource for researchers to use in assessing and improving the accuracy of lesion segmentation tools.

304 MRI images from 11 cohorts worldwide were collected from research groups in the ENIGMA Stroke Recovery Working Group consortium. Images consisted of T1-weighted anatomical MRIs of individuals after stroke. These images were collected primarily for research purposes and are not representative of the overall general stroke population (e.g., only including individuals who opt in to participate in a research study, and excluding individuals with stroke who cannot undergo MRI safely).

3a8082e126