Skip to Content
DEEPCRAFT™ Studio 5.10 has arrived. Read more →

Release Notes 5.10

This section lists the new functionality, improvements and some of the major changes related to DEEPCRAFT™ Studio.

New Features and Enhancements

New tools and functionality for image labeling

New tools, features, and keyboard shortcuts are now available for creating and managing bounding boxes in Object Detection projects, enabling faster and more efficient labeling.


  • Copy and paste bounding boxes: Copy one or more bounding boxes and paste onto the same image, or paste across multiple images using the Select tool and keyboard shortcuts.
  • Zoom and pan for precision: Zoom in for detailed labeling, zoom out for a broader view, and navigate the image using the Pan tool.
  • Toggle overlays: Hide or show bounding boxes and model prediction overlays using the eye tool.
  • Delete: Delete multiple bounding boxes on the same image using the Delete tool.
  • Undo or redo support: Undo and redo annotation actions, including deleting, moving, and resizing bounding boxes.

To know the detailed instructions, refer to Labeling Images

New unit nodes in GraphUX

Graph UX now includes the following nodes, available in the Node Explorer. Each unit provides implementations for float32, int8, and uint8.

  • Trigonometry: Sin, Cos, Tan, Sinh, Cosh, Tanh, ArcSin, ArcCos, ArcTan, Deg2Rad, Rad2Deg

  • Arithmetic: Division (element-wise with modulo mode)

  • Utilities: Cast unit supports type conversion with Clamp/Range modes. The unit also supports conversions between float32, signed integers (int8, int16, int32), and unsigned integers (uint8, uint16, uint32).

  • Math: Argmin which returns the index of the minimum value along a selected axis (complements the existing Argmax node).

  • Machine Learning: Consecutive Detections Filter, which reduces false positives by requiring class confidence to remain above a threshold for N consecutive samples.


Model Config tab in .h5 file

A new Model Config tab is now available for .h5 models. This tab provides an at-a-glance view of the different training configurations used to train the model.


Fixes

  • Overall bug fixes and increased stability.
  • Performance improvement for searching nodes in Graph UX
  • Performance improvement for Data Tab functionality
Last updated on