DIVE is an open source rapid annotation and AI training interface for video and image data. Kitware developed DIVE so that users have the ability to easily generate annotations, whether it be manually using interactive web editing tools or programmatically using AI models. It uses the latest AI models that can pre-populate annotations, saving intensive labor hours of manually creating annotations from scratch. DIVE includes the latest UI/UX best practices to enable fast annotation creation and editing for domain scientists.
Key features of DIVE
- Video annotation creation and editing
- Still image (and image sequence) annotation creation and editing
- Video and image data and metadata management
- Deep integration with VIAME computer vision analysis tools
- Single-frame boxes, polygons, and lines
- Multi-frame bounding box tracks with interpolation
- Automatic transcoding to support most video formats
- Customizable labeling with text, numeric, multiple-choice attributes
- Multiple dataset training
DIVE is typically packaged within the VIAME toolkit, Kitware’s open-source do-it-yourself AI system. However, it can also be used as a standalone application or can be integrated into other analytics platforms. Due to its connection to VIAME, DIVE can also enable training new models stored within the VIAME repository using existing annotations. DIVE’s web- and cloud-based hybrid workflow enables training jobs to run on separate machines in or outside the network, such as Google cloud.
DIVE can be web or desktop-enabled
DIVE can run on local servers or the cloud as a web service using DIVE’s container images. You can view a sample instance of DIVE running on a public server here.
We are pleased to announce that DIVE is also available as a desktop application. This desktop version brings most of the features of the web version for standalone use without requiring a network connection or a server installation. This is specifically useful for those who are working in remote locations with limited internet access, or users who are working on isolated efforts. The desktop version is fully supported on Windows, macOS, and Linux. You can download DIVE Desktop here.
The Experts Behind DIVE
Our Data and Analytics Team has created a platform of tools to enhance the way customers use data. These tools have proven valuable across multiple disciplines from scientific research, to engineering design, to software development. With scalable, web-based data management, powerful, intuitive analytics engines, and functionality to create graphs, charts, geospatial maps, and other visualizations, these tools are designed to enable powerful custom solutions for any data need. We partner with universities on government-funded projects and with commercial companies on product development. Learn more about the capabilities of our Data and Analytics Team by visiting Kitware’s data and analytics page.
Our Computer Vision Team has been a leader in deep learning for computer vision since 2014. Our technology centers on state-of-the-art automated image and video analysis, including in the areas of DIY and ethical AI. Our DIY AI toolkits give users the ability to train object classifiers using interactive query refinement and to improve existing capabilities to work optimally on your data for tasks such as object tracking, object detection, and event detection. Our toolkits also allow you to perform customized, highly specific searches of large image and video archives powered by cutting-edge AI methods. If you would like more information about our Computer Vision Team, please visit Kitware’s computer vision page.
Leveraging DIVE for Your AI Needs
DIVE is open source and can be used without restriction, but partnering with Kitware will ensure that you are using the application effectively and getting the results you’re looking for. We can work closely with you to develop a custom solution tailored to your specific requirements. If you’re interested in using DIVE for your AI project, contact us at email@example.com.
The Kitware team recently presented a DIVE tutorial to the National Oceanic and Atmospheric Administration (NOAA) team during a 5-day tutorial workshop. We will be sharing more details on this in an upcoming blog.
This work is supported by the Intelligence Advanced Research Projects Activity (IARPA) via contract 2017-16110300001. The U.S. Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright annotation thereon. Disclaimer: The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of IARPA or the U.S. Government.
VIAME was originally developed with funding from the National Oceanic and Atmospheric Administration (NOAA)’s Fisheries Strategic Initiative on Automated Image Analysis (AIASI) committee, in addition to other groups within NOAA, such as the Marine Mammals Lab, and external groups such as the Coonamessett Farm Foundation.