Semantic Segmentation | Tri Geo Technologies

E:info@trigeo.ai

Semantic Segmentation

In Semantic Segmentation pixels of an image needs to be linked to a labeled image, for example a car, bike, traffic lights or pedestrian. It is further used to segment objects of a similar class by giving unique id’s to every object.

Capabilities

Semantic Segmentation Use Cases :

Instance segmentation for feature detection - Used in training perception models in non-environmental objects of interest.

3D Semantic Segmentation - Get highly accurate segmented 3D point clouds with support for vast number of classes. We support distinctions between instance level segmentation, drivable and pedestrian designated zones. Unique features like point size controller, ground and ceiling adjuster, segmenting using polygons enable us to provide accurate point segmentation.

Full pixel semantic segmentation - Highly utilized in autonomous vehicles where the information of every pixel is highly critical and it may influence the accuracy of the perception model.