New updates and improvements
V7 Darwin

Before the update, Single-select and Multi-select properties had to be applied to specific annotations and classes, like a bounding box. Now, you can add them independently as Item Attributes—custom fields that apply to entire files or groups of files, rather than specific regions. This feature helps you capture higher-level insights and contextual information.

With Item Attributes, you can:

• Add Single-select, Multi-select, or Text fields that apply to whole files

• Make specific fields required to ensure data completeness

• Display attributes and questions in the annotation panel

• Capture information that spans across multiple files

To add Item Attributes, go to the Classes tab in your dataset. Set up Single-select, Multi-select, or Text type attributes and write questions/instructions for annotators. Additionally, it is possible to make attributes required before moving files to the next workflow stage.

Learn more: https://docs.v7labs.com/docs/item-attributes

Many V7 Darwin users have shared that they prefer the UX of our proprietary Auto-Annotate model over point-based segmentation. We've redesigned our SAM 2 implementation to make object segmentation more efficient and to give you the choice between both modes. You can now switch between Point and Click and Box modes for our segmentation tool powered by SAM.

Key improvements:

• Draw a quick box around objects for instant segmentation

• Get accurate results without placing multiple points

• Focus on the object you want, not the background you don't

• Full compatibility with both SAM 2 and V7's Auto-Annotate models

Instead of relying on multiple positive and negative points for selection, you can now simply draw a box around your target object to get precise automatic segmentation. For more complex cases, you can still refine results with additional point-based adjustments, but most objects can now be isolated with a single box selection.

In video annotation, some properties of an object may be dynamic and vary throughout the timeline—for example, a vehicle, tool, or person might change position, orientation, or state as they move through the scene. However, other attributes, such as its material, brand, size, or type, are unlikely to change. With this update, you can now set specific properties to be either dynamic or permanently static.

Use this feature to:

• Define properties that remain constant or change throughout the clip

• Mix static and keyframe-based properties on the same annotation

• Prevent accidental property changes across frames

Keyframe enabled/disabled properties ensure that fixed values stay consistent across the entire video. This prevents accidental edits and helps you maintain accuracy for attributes, whether they vary or not.

Learn more: V7 Darwin Documentation

We've upgraded our annotation panel to show single and multi-select properties. Now you can instantly spot annotation types and sub-types. This feature makes training data review much easier and more intuitive, even with complex class taxonomies.

The update is particularly valuable for teams incorporating human-in-the-loop and RLHF techniques in their machine learning development. It simplifies the work of data annotators and improves overall efficiency in MLOps workflows.

Channels allow you to overlay and cycle through multiple visual layers, making it easier to annotate complex data. This feature is particularly useful for annotation in fields such as microscopy, digital pathology, remote sensing, or geospatial imaging.

To get started, import your files via our API, specify a base channel for annotations, and access Channels via the sidebar menu or by pressing T. Future enhancements will include support for opacity adjustment controls.

We've introduced a new highly requested annotation feature for CT, MRI, and other volumetric data. You can now use a 3D brush to create voxel masks and apply them to regions of interest based on brightness intensity.

The feature includes:

• Intelligent thresholding based on pixel intensity

• 3D sphere and cube brush shapes with adjustable sizes

• Consistent segmentation across all MPR planes

To enable thresholding, click on the 3D brush settings or use the Ctrl + T shortcut. You can also use the image manipulation panel and apply levels to find the correct values. For the most convenient and effective use of this tool, adjust the window levels to set brightness thresholds and then automatically match the settings to your brush.

Learn more: V7 Darwin Documentation

We’ve integrated the latest version of Meta’s computer vision segmentation model into V7. The updated Segment Anything Model 2 (SAM 2) offers improved accuracy for segmenting complex objects with minimal input. It is also more suitable for video annotation. Very often, you need just a single click to track objects throughout an entire video sequence.

This implementation includes:

• More accurate segmentation and easy adjustment

• Consistent movement tracking in clips and longer videos

• Improved model performance for domain-specific scenarios

The model can maintain context about objects even when they temporarily disappear, which is very valuable for challenging real-world scenarios. This functionality also pairs perfectly with V7's In and Out of View detection.

To get the most out of SAM 2, you can also use the timeline to adjust selected frames and recalculate auto-tracking for the selected section of the clip. This powerful addition expands SAM 2's capabilities beyond the out-of-the-box functionalities the model offers.

Learn more: https://ai.meta.com/blog/segment-anything-2/

Focus Mode maximizes your annotation screen space by hiding non-essential UI elements. In fields like medical imaging, where pixel-perfect accuracy is critical, Focus Mode ensures that every inch of your monitor is used effectively.

Activate Focus Mode by pressing ⇧Shift + F to turn on a distraction-free, full-screen annotation panel optimized for precise work.