Camera-based in-cabin object localization

A method for performing object localization inside a cabin of a vehicle, comprising:

• maintaining a machine-learning model trained on images taken at a first location within a cabin of a vehicle, the vehicle having a seating zone configuration having a plurality of seating zones, the machine-learning model configured to place one or more objects into the seating zones according to a plurality of bounding boxes corresponding to the plurality of seating zones from a vantage point of the first location;
• receiving a camera image taken of the cabin of the vehicle by an image sensor at a second location with respect to the plurality of seating zones of the vehicle;
• projecting the plurality of bounding boxes based on a geometry defined by the first location and the second location to adapt the machine-learning model to a vantage point of the second location;
• performing object detection on the camera image to identify the one or more objects in the camera image;
• and
• utilizing the machine-learning model trained on the images taken at the first location to place the one or more objects into the seating zones of the vehicle according to the projected plurality of bounding boxes corresponding to the plurality of seating zones from the vantage point of the second location.

A non-transitory computer-readable medium comprising instructions for performing object localization inside a cabin of a vehicle that, when executed by one or more computing devices cause the one or more computing devices to perform operations including to:

• maintain a machine-learning model trained on images taken at a first location within a cabin of a vehicle, the vehicle having a seating zone configuration having a plurality of seating zones, the machine-learning model configured to place one or more objects into the seating zones according to a plurality of bounding boxes corresponding to the plurality of seating zones from a vantage point of the first location;
• receive a camera image taken of the cabin of the vehicle by an image sensor at a second location with respect to the plurality of seating zones of the vehicle;
• project the plurality of bounding boxes based on a geometry defined by the first location and the second location to adapt the machine-learning model to a vantage point of the second location;
• perform object detection on the camera image to identify the one or more objects in the camera image;
• and
• utilize the machine-learning model trained on the images taken at the first location to place the one or more objects into the seating zones of the vehicle according to the plurality of bounding boxes corresponding to the plurality of seating zones for the first location.