Machine Learning

Discover Metavision ML, a comprehensive Machine Learning framework for event-based vision.

Metavision ML uses PyTorch as a machine learning framework. This gives us the power of a well established framework, the convenience of Python and C++ integration, and the possibility to export your event-based model to other frameworks.

Event-based data is typically temporally rich and spatially sparse, which makes it difficult to exploit using the currently available machine learning frameworks. In Metavision ML module, we ensure that event-based data is correctly managed so training loops and inference pipelines are designed to guarantee the compatibility between event-based data and generic machine learning frameworks.

To be used in modern deep learning models, which use tensors to accelerate the training, our raw events are encoded into tensors. This step is detailed in Event Preprocessing Tutorial

Therefore, with Metavision ML you can use the advantages of event-based vision together with the power of machine learning. Currently, we provide pre-trained models and inference samples for various applications like:

• detection and tracking

• optical flow

• classification

• corner detection

In addition, you will find all the tools required to ensure the best use of event-based data for machine learning:

• tools to pre-process data to create tensors

• training loops for object detection, classification, and optical flow

• conversion tools to create advanced and information-rich images from event-based data

• large-scale datasets

• tools for dataset creation: labeling, visualization, and evaluation