eCognition 10.2.1 is available!
The latest version of eCognition 10.2.1 is available! It includes improved usability of the deep learning algorithms for the users with insufficient GPU. eCognition now offers a new parameter that is available in the algorithm ‘set rule set parameters’. The parameter is called ‘Deep Learning CPU lock’ and it enables more flexible options for the hardware resources utilization. Previously, when using deep learning algorithms (‘apply instance segmentation’, ‘apply convolutional neural network’ etc.), users with poor GPU resources encountered error messages and had to follow a complicated workaround to be able to use these algorithms. With this new option, we make it possible to easily switch to the CPU usage.
We made a few UI improvements related to the creation of applications for their further usage in eCognition Architect, including slider widget enhancement and minor fixes in the dialog windows.
We also fixed a number of important bugs. Now you can expect the algorithm ‘remove objects’ and the option ‘Save Ruleset minimized’ to work as expected. More minor fixes are included in this version, the details can be found in our Release Notes.
Was this article helpful?