Index

Encord Index empowers AI & ML teams to efficiently create high quality, balanced datasets to improve AI model performance at scale without compromising speed or security. With Encord Index you can unify multimodal data from all local and cloud data sources to one platform for in-depth data visualization, search and granular curation.

​
When to use Encord Index?

Encord Index helps you visually understand and improve your data at all stages of your machine learning journey. Index does not just store your data, you can curate your data too.

Before going any further, you should know what a Collection is in Encord Index. Collections provide a way to save interesting groups of data units to support and guide your downstream workflow. For more information on Index Collections go here.

​
What data does Encord Index support?

Index supports analysis (Advanced Metrics and Embeddings) on images and videos up to 4K resolution. Performance is affected for images and videos over 4K.

For optimal performance, we strongly recommend downscaling images and videos over 4K to 4K resolution. Videos import to Index at 1 FPS. You can specify specific frames you want imported from your videos. Contact us for more information on importing specific video frames to Index.

FeaturesIndex
Image typesjpg, png, tiff
Video typesmatroska and webm (vp8, vp9, libdav1d), mov, mp4, m4a, 3gp, 3g2, mj2 (libdav1d, h264, h265)
Audio typesmp3, .mp4, .wav, .m4a, .flac, .eac3
Number of images10,000,000 per Folder (unlimited Folders)
Videos *2 hours @ 30fps
Data exploration✅
Similarity search✅
Off-the-shelf quality metrics✅
Image duplication detection✅
Outlier detection✅
Collections✅
Integration with Encord Annotate✅
Custom metadata✅
  • Contact Encord to discuss optimal strategies for indexing longer videos.

Was this page helpful?

Video TutorialsGetting Started with Index