Dense 6-DoF pose object detector

1. A computer-implemented method for object detection and pose estimation within an input image, the method comprising:

receiving the input image;

inputting the received input image into a trained encoder-decoder convolutional artificial neural network comprising an encoder head, an ID mask decoder head, a first correspondence color channel decoder head, and a second correspondence color channel decoder head;

generating, using the ID mask decoder head, an ID mask identifying one or more objects and background in the received input image;

generating, using the first correspondence color channel decoder head, a first correspondence color channel of a 2D-to-3D-correspondence map for the one or more objects within the received input image;

generating, using the second correspondence color channel decoder head, a second correspondence color channel of the 2D-to-3D-correspondence map;

generating the 2D-to-3D-correspondence map using the generated first correspondence color channel and the generated second correspondence color channel; and

determining, for at least one object of the one or more objects identified by the ID mask, a pose estimation based on the generated 2D-to-3D-correspondence map and on a pre-generated bijective association of points of the object with unique value combinations in the first correspondence color channel and the second correspondence color channel.

2. The method of claim 1 , wherein the first correspondence color channel is a Blue color channel, wherein the second correspondence color channel is a Green color channel, or wherein the first correspondence color channel is the Blue color channel and the second correspondence color channel is the Green color channel.

3. The method of claim 1 , wherein determining of the pose estimation uses a Perspective-n-Point (PnP) algorithm.

4. The method of claim 3 , wherein the PnP algorithm is used with a Random Sample Consensus (RANSAC) algorithm.

5. The method of claim 1 , wherein determining of the pose estimation uses a trained artificial neural network entity configured and trained to generate, from the ID mask and the 2D-to-3D-correspondence map, probabilities for each of a plurality of 3D poses of the at least one object.

6. The method of claim 1 , further comprising:

generating the pre-generated bijective association for at least one object, the pre-generated bijective association generated by texturing a 3D representation of an object using a 2D correspondence texture consisting of a plurality of pixels, each pixel including a unique value combination in the first and the second correspondence color channels.

7. The method of claim 6 , wherein texturing the 3D representation of the at least one object is performed using a spherical projection.

8. A System for object detection and pose estimation within an input image, the system comprising:

an input interface for receiving the input image;

a computing device configured to implement a trained encoder-decoder convolutional artificial neural network comprising an encoder head, an ID mask decoder head, a first correspondence color channel decoder head and a second correspondence color channel decoder head;

wherein the ID mask decoder head is configured and trained to generate an ID mask identifying one or more objects and background in the received input image;

wherein the first correspondence color channel decoder head is configured and trained to generate a first correspondence color channel of a 2D-to-3D-correspondence map for the one or more objects within the received input image;

wherein the second correspondence color channel decoder head is configured and trained to generate a second correspondence color channel of the 2D-to-3D-correspondence map;

wherein the computing device is further configured to implement a combining module and a pose determining module;

wherein the combining module is configured to generate the 2D-to-3D-correspondence map using the generated first correspondence color channel and the generated second correspondence color channel; and

wherein the pose determining module is configured to determine, for an object of the one or more objects identified by the ID mask, a pose estimation based on the generated 2D-to-3D-correspondence map and on a pre-generated bijective association of points of the object with unique value combinations in the first correspondence color channel and the second correspondence color channel.

9. The system of claim 8 , wherein the first correspondence color channel is a Blue color channel, wherein the second correspondence color channel is a Green color channel, or wherein the first correspondence color channel is the Blue color channel and the second correspondence color channel is the Green color channel.

10. The system of claim 8 , wherein the pose determining module is configured to determine of the pose estimation using a Perspective-n-Point (PnP) algorithm.

11. The system of claim 10 , wherein the PnP algorithm is used with a Random Sample Consensus (RANSAC) algorithm.

12. The system of claim 8 , wherein the pose determining module is configured to determine the pose estimation using a trained artificial neural network entity configured and trained to generate, from the ID mask and the 2D-to-3D-correspondence map, probabilities for each of a plurality of 3D poses of the object.

13. The system of claim 8 , wherein the computing device is further configured to generate the pre-generated bijective association for at least one object of the one or more objects, the pre-generated bijective association generated by texturing a 3D representation of an object using a 2D correspondence texture consisting of a plurality of pixels, each pixel including a unique value combination in the first and the second correspondence color channels.

14. The system of claim 13 , wherein texturing the 3D representation of the at least one object is performed using a spherical projection.