Roboflow Changelog: May & June 2022

Welcome to a Universe (opens in a new tab) sized edition of the Roboflow Changelog! Each month, we share a recap of product and company updates; if you missed the last edition of the Roboflow Changelog, you can find it here.

A note on licensing: YOLOv5 is released under AGPL-3.0, which affects commercial use:

• On Roboflow cloud: a commercial license is included for every user, including the free plan.
• Self-hosted: you need an Enterprise license add-on, or you must open-source your application under AGPL-3.0.
• Permissive alternative: RF-DETR (opens in a new tab) is Apache 2.0, free for commercial use anywhere.

Roboflow Organize and Annotate

• Launched a new role for only labeling images
• Added Oriented bounding boxes, YOLOS, and YOLOv6 support
• Enabled Google SSO for login
• Added the ability to rename and delete batches and jobs (opens in a new tab)
• Labelbox Video Import
• Added automated active learning to roboflow the pip package

Roboflow Train and Deploy

• Launched Native iOS SDK
• Launched OAK pip package
• Launched Instance Segmentation Train & Deploy
• Polygon and box improvements to improve performance
• Improved edge inference tiling

Roboflow Universe (opens in a new tab)

• Released V2 of Roboflow Universe (opens in a new tab): The world's largest resource of computer vision datasets and pre-trained Models!
• Added a Model tab so users can quickly try and use any pre-trained model
• Created and launched tags and aggregate pages (opens in a new tab)
• Added video, webcam and URL options to inference widget
• Enhanced search capabilities
• Added advanced health check to projects
• Enabled edge deploy of Universe projects

Other

• Microsoft added Roboflow to the Azure Marketplace (opens in a new tab)
• Created a new Knowledge Base (opens in a new tab)
• Onboarding enhancements
• Integrated Salesforce
• Fixed 24 bugs
• Team meet up in San Francisco
• Roboflow hosted a recruiting event in Austin
• Celebrated 1 year anniversaries for 2 team members
• Added 2 new team members
• Opened several positions (come help us democratize computer vision!)
• Conducted UXR sessions
• 59 discussions on the Roboflow Forum (opens in a new tab)
• 10k+ new Universe projects (opens in a new tab)
• 17 new blog posts
• 3 new YouTube videos (opens in a new tab)
• Featured in/by: VentureBeat (opens in a new tab), LearnOpenCV (opens in a new tab), Rising Above the Deluge (opens in a new tab), Bosch Future Mobility Challenge (opens in a new tab), Virginia Tech (opens in a new tab), University of Manchester (opens in a new tab), Soar (opens in a new tab), Flex Logix (opens in a new tab), Native Digital (opens in a new tab), Azure Marketplace (opens in a new tab),
• User projects: TUM Autonomous Motorsport: An Autonomous Racing Software for the Indy Autonomous Challenge (opens in a new tab), A System for a Real-Time Electronic Component Detection and Classification on a Conveyor Belt (opens in a new tab), Build Android app for custom object detection (TensorFlow 2.x) (opens in a new tab), License Plate Image Threshold Filtered (opens in a new tab), Fire and Smoke Realtime Detection System (opens in a new tab), DropTrack - automatic droplet tracking using deep learning for microfluidic applications (opens in a new tab), Detecting numbers on a racing bib in a natural scene (opens in a new tab), ViDMASK dataset for face mask detection with social distance measurement (opens in a new tab), Skin Disease Diagnosis Using Image Analysis and Natural Language Processing (opens in a new tab), Bib Detection (opens in a new tab), Copter Spotter Vision (opens in a new tab), Sign Language Recognition from a webcam video stream (opens in a new tab), VineInspector: The Vineyard Assistant (opens in a new tab), Automatic Classification of Hospital Settings through Artificial Intelligence (opens in a new tab), Few-Shot Object Detection with YOLOv5 and Roboflow (opens in a new tab), Object Detection for Inventory Stock Counting Using YOLOv5 (opens in a new tab), Smart Quality Management System at the Manufacturing Sector using Deep Learning for Anomaly Detection (opens in a new tab), Object Detection at the Edge with TF lite model-maker (opens in a new tab)