What is annotation?

Annotation is the process of tagging or labelling various forms of data or content such as images, text, audio,or video. This is done for machine learning models to recognise them and in addition make predictions.

What is autonomous vehicle annotation?

Vehicle annotation is done by using  boundary boxes and defining other attributes which train machine learning models to recognise and understand objects detected by the sensors in the vehicle.

Types of annotation for autonomous vehicles 

2D boxing annotation for self-driving cars

2D boxing annotation, which is rectangular,makes use of computer vision to train autonomous vehicles in recognising objects while on the streets. These include street lanes, road signs, parking zones and vehicles.  It is the simplest form of annotation for autonomous vehicles. 

3D Point Cloud  annotation for autonomous vehicles 

This type of data training allows for precise  object detection with the use of sensors such as LiDAR, which transmit light and calculate the time it takes for reflection back into the sensor for each point to be created with 3D boxes.  It makes both indoor and outdoor objects recognisable. For automated vehicles, it is used for distinguishing and classifying lanes on roads with the use of 3D point cloud maps

Semantic segmentation for autonomous vehicles 

Semantic segmentation is the digital technique of dividing or partitioning an image into various parts or regions, taking into account the image's pixels. With automated vehicles,annotated objects are shaded to be easily recognised through computer vision.Semantic segmentation differentiates all the content in the image.

Polygon annotation for self-driving vehicles 

Polygon annotation automates the detection of  complex-shaped objects which are in high demand for accuracy.  It draws precise polygons around objects with odd shapes. It helps automated vehicles to recognize visible objects such as motorcycles,bicycles or cars on the street. Even though it is very accurate,it is also costly and very time-consuming.

Key point annotation for autonomous vehicles 

This kind of data training makes streets and highways easily recognised for accurate road movements. It makes use of computer vision to annotate road surfaces and lanes (whether single, double or broken, painted ones) for easy detection by automated vehicles.

It features a varied number of connected and numerically labelled edges called key points. They are essential for tracking variations in similar structured objects such as the human anatomy. For autonomous vehicles, key points are used as touch points between roads and vehicle tyres.

Polyline annotation for autonomous cars

This kind of data training makes streets and highways easily recognised for accurate road movements. It makes use of computer vision to annotate road surfaces and lanes(whether single, double or broken, painted ones) as well as pedestrian crosswalks for easy detection by automated vehicles.

Single  and streamed video annotation for autonomous vehicles 

With this technique, videos are divided into thousands of images while targeted images are annotated for every single frame.They essentially help vehicles to recognise objects in complex scenarios.Streamed frame video annotation uses algorithms from objects annotated in single frame videos to track targeted objects in subsequent frames. Annotations are only adjusted in cases where  algorithms are unable to function to satisfaction. It is used for simple scenarios .

‍