Aivia That Learns
Faster neuron detection with AI
Aivia’s AI image analysis capabilities continue to grow with the enhancement of Aivia That Learns, an AI-powered autonomous parameter prediction tool. This feature, introduced in Aivia 10, reduces the need for a time-consuming manual parameter setup.
In this release, we add a deep learning model which can autonomously set all the image analysis parameters for 3D neuron detection and tracing in fluorescence images. Accessed through Aivia’s user-friendly Active Tiles navigation bar or within the Recipe Console, Aivia That Learns now empowers you to kick off your neuron tracing analysis with just two mouse clicks.
-
Optimal analysis parameter prediction for 3D neuron images
-
Easy access via Active Tiles or the Recipe Console
-
Optional feedback system to help improve Aivia That Learns
You can also help us improve Aivia That Learns for the benefit of the entire neuroscience community. Aivia includes an opt in user-data gathering system so that you to easily share representations of your images, recipe parameters and feedback on the quality of the analysis with our development team. We will use this data to update future version of Aivia That Learns for better parameter prediction for everyone.
Aivia Elevate for Neuro and Apex subscribers have automatic access to this feature. Learn more about the Aivia subscription plans here.
Based on our testing, you can shorten the time from raw image to fully analyzed data set by up to 74%.
Media Gallery
Deep Learning Expert Mode
Boost your deep learning capabilities
Aivia 10.5's AI updates do not stop with Aivia That Learns. In this release, we have also expanded the functionalities of our deep learning pipeline.
Built for expert users who require more freedom and flexibility, it is now possible to set all train and apply hyperparameters. Choose the optimal network architecture, intensity normalization method, learning rate, number of epochs and much more for your deep learning model training and applying.
-
Customize your deep learning hyperparameters for training and applying
-
Supports new TensorFlow, CUDA and Python versions
-
Covering more GPUs
We have also implemented a new deep learning JSON based file format (.aiviadl) so you can easily open, read and understand the hyperparameters used to create your models. On top of this, we have updated TensorFlow, CUDA and Python as well as extended support for recently released NVIDIA GPUs.
With these new and improved features, you can achieve more with Aivia’s innovative AI-powered image enhancement and analysis tools.
Media Gallery
Streamlined User Experience
More tools at your fingertips
Deploying Aivia’s innovative image analysis tools is now more convenient than ever. Aivia 10.5’s user interface includes smart Active Tiles that adapt to your preferences and a new search bar functionalities that gets you to where you need to be faster.
You can also get the most out of your Leica and Aivia systems with extended support for STELLARIS 8 FALCON files such as tilescan Thunder processed ICC images. Spend less time converting files and more time viewing and analyzing your images.
-
User-driven Active Tile display
-
More tools available in the search bar
-
Easily load and open your STELLARIS 8 FALCON images
