July 2023 Summaries

This week's FiftyOne tips and tricks blog covers various topics such as listing saved views in a dataset, exporting datasets to different formats, handling tags in the FiftyOne App, importing datasets with custom formats, managing FiftyOne Brain keys, and uploading images with nested directories. The community is encouraged to join the FiftyOne Slack channel for more discussions and support.

This week's FiftyOne tips and tricks blog covers interesting questions and answers from the community. Topics include more effective source code searches, renaming semantic mask paths, working with cloud-based notebooks using proxy_url, exporting and sharing dataset views, making label boxes invisible, and an introduction to FiftyOne Teams for collaboration. The blog also encourages readers to join the growing FiftyOne community on Slack, GitHub, Meetup, and more.

The July 20, 2023 Computer Vision Meetup featured presentations on various topics such as learning new dimensions of human visual similarity using synthetic data, self-supervised approaches for facial representation learning from videos, and the role of vector databases in computer vision applications. Attendees also had the opportunity to vote for their favorite charities, with donations being made to Drink Local Drink Tap and Education Development Center. The next Computer Vision Meetup is scheduled for August 10, featuring speakers on neural congealing, advancing personalized medicine through AI-enabled computer vision, and a practical approach to deep learning for computer vision with Tensorflow 2.

The July 13, 2023 Computer Vision Meetup featured discussions on various topics such as vector databases in computer vision applications, semantic search at scale over unstructured data objects, reverse image search for e-commerce, and scalable image and text search. The event also highlighted the role of FiftyOne and Milvus vector databases in healthcare, retail, finance, and more. Additionally, it covered how to build scalable image and text search using Pinecone and Qdrant. The Meetup aimed at bringing together communities of data scientists, machine learning engineers, and open-source enthusiasts who want to share and expand their knowledge of computer vision and complementary technologies.

Vector search engines are becoming essential components in AI applications, transforming complex data into numerical vectors that can be efficiently searched against raw data. However, working with image or video data and incorporating vector search into workflows can present challenges such as implementing cross-modal retrieval and handling traditional search filters. FiftyOne is an open source computer vision interface for vector search, allowing users to use their preferred vector search engine in a single line of code. With native integrations with Qdrant, Pinecone, LanceDB, and Milvus, FiftyOne simplifies semantic searches using natural language queries and supports object detection patches.

The FiftyOne blog series highlights recent developments in the FiftyOne community and recognizes significant milestones. Community spotlights feature various use cases of FiftyOne, such as object tracking inside warehouses by Qinecsa and leaf-level insights for farmers by Taranis. Users can share their success stories with FiftyOne to receive a box of community rewards. The latest product release is v0.21.1, which includes App optimizations and several enhancements. Community contributors are acknowledged for their work on the project. The FiftyOne GitHub repository has over 3,800 stars, 63 contributors, and 371 forks. The FiftyOne Community Slack channel connects more than 1,800 machine learning engineers and data scientists who share tips and tricks in a weekly blog series. Voxel51 sponsors 13 virtual Computer Vision Meetups worldwide, with upcoming events scheduled for July and August. New resources include blogs, videos, tutorials, integrations, and cheat sheets to help users succeed with FiftyOne.