Robovision AI 5.4 release notes¶

Release date: February 14, 2024

What's new¶

In the thumbnails view of the label center, you can now easily select multiple samples by:
- Dragging over samples.
- Holding down the Ctrl/Cmd key while clicking individual samples.
- Holding down the Shift key and selecting the first and last samples.
In the thumbnails view of the label center, you can make sample thumbnails even smaller to view more samples at once.
You can now filter samples by a source—that is, members who labeled the samples or prediction jobs that generated the labels.

The project details page has been redesigned for enhanced usability and support of custom jobs defined by certain project types.

Previously, members who were created using SDK or API had the Not accepted status and were not required to read or accept the terms and conditions. They will now be prompted to do this during their next attempt to sign into the Robovision AI.
When creating a new member, you can now assign them one of these roles: Administrator, Labeler, or Data scientist. See Member roles.

Settings page has been renamed to System and split into these sections:
- About – view platform information and logs.
- Members – manage platform users.
- Resources – easily monitor running processes, GPU, and memory usage.
Running jobs have been renamed to Running processes. This change is purely cosmetic, and the underlying concept and functionality remain unchanged.

Robovision AI now supports upgrading Robovison Edge. When you start the upgrade procedure in Robovison Edge, the upgrades are fetched from Robovision AI.

The issue with platform erroring when navigating through the platform has been fixed.
On the Running processes page, a process was shown as running even if it had no resources. This has been fixed.

Previously, anyone could edit an annotation—now, this can be done only by its owner.
Confusion matrix showed testing predictions as labels of an Unknown user. This has been fixed, and testing predictions are now hidden from the charts.

If training data contains empty annotations, they are now shown in the class distribution to provide an accurate overview of data distribution.

Previously, running models could be selected for testing in the test center but not on the model comparison page. Now, you can test models from both locations.
Tests that were run with enabled GPU failed on Azure deployments. This has been fixed.

The Robovision AI UI has been optimized for the browser size of 1920x1200.
You will be logged out of the platform if your data import has been in the Uploading stage for more than an hour.
If the storage is full, some functions may become unavailable, for example, saving annotation or training a model.
Renaming a project while there is an ongoing import process may cause the import process to restart.
When upgrading to a new version, new algorithm versions are installed, while preserving the existing ones. Existing projects will be updated to utilize the new algorithms.
The class distribution chart in the label and test centers doesn't consider the filters you have applied—the number and percentage of samples labeled with a certain class is incorrect. Click the needed bar to view the actual number of samples.

The Robovision AI platform has been designed and validated for Google Chrome 85 and Mozilla Firefox 77 or later.
In Google Chrome, training metrics and a final validation score might not be displayed if an ad blocker is used. To prevent this, mark the Robovision AI URL as trusted in the Google Chrome settings.
In Google Chrome, the thumbnail view can load slowly if you have password managers like Lastpass activated. To prevent this, disable the extension or work in a guest profile window.

Backward compatibility with Robovision AI versions 3.x and 4.x is not foreseen. But upon request, Robovision will provide the needed assistance and conversion scripts.

Types supported: In general and with respect to the model testing functionality, the following image types are supported: JPEG, JPG, PNG, TIF, TIFF, BMP, GIF, including a 4^th transparency channel. EXIF orientation metadata is supported.
File names must contain only Latin (ASCII) characters. There are known rendering, packing, and backup/restore issues with file names/object keys that contain non-Latin characters.

2D labeler performance (in particular the Magnetic lasso and Grab cut tools) may degrade for high-resolution images.
You may face the following issues when using the Prediction tool in the label center:
- The prediction process gets hidden if you leave the label center. To view it again, click the Prediction tool.
- The predicted labels are not added to the samples as the prediction process is running.
- The Status section is not updated as the prediction process is running.
- Multi-Label EfficientNet, SOLOv2, and YOLOv5: If you delete an annotation generated by the prediction tool, the credibility score for all other annotations generated by the prediction tool for the same sample will disappear.
You can copy or replace annotations for a maximum of 100 samples at a time.
In the label center, it's not possible to delete annotations from more than 1000 samples.

During training setup, the Add model button to select a model for transfer learning is enabled even if there is no model available.
It's possible to start a training with test predictions even if the project doesn't contain annotated data or datasets.

Multi-Label EfficientNet: The Confidence threshold parameter is not displayed during the inference setup.

You won't receive notifications about data upload from Robovision Edge.
On the project details page, you can't stop the data upload from Robovision Edge.