Method for generating synthetic data

1. A method for generating 2D images, the method being carried out by a processing unit and comprising the steps of:

generating a synthetic 3D model of an object with annotation points;

based on the synthetic 3D model of the object, generating a plurality of synthetic objects with different poses and different brightness conditions;

receiving a plurality of 2D images with different backgrounds and different brightness conditions;

calculating a brightness value for each of the plurality of 2D images; and

for each of the plurality of 2D images, inserting one of the plurality of synthetic objects based on the corresponding brightness value.

2. The method according to claim 1 , wherein inserting the one of the plurality of synthetic objects comprises randomly inserting the one of the plurality of synthetic objects.

3. The method according to claim 1 , wherein inserting the one of the plurality of synthetic objects comprises, for each of the plurality of 2D images, inserting the one of the plurality of synthetic objects that presents the same brightness conditions as a corresponding one of the plurality of 2D images.

4. The method according to claim 3 , wherein the brightness conditions comprise an intensity of a background of the corresponding one of the plurality of 2D images.

5. The method according to claim 3 , wherein inserting the one of the plurality of synthetic objects comprises:

generating the plurality of 2D images from video acquired using a 2D camera;

modeling an environment in which the 2D camera operates to acquire the plurality of 2D images with the brightness conditions; and

associating a synthetic object of the plurality of synthetic objects each one of the plurality of 2D images having same brightness conditions.

6. The method according to claim 1 , wherein inserting the one of the plurality of synthetic objects comprises:

based on the brightness values, associating a 2D image of the plurality of 2D images with a hand image having a same brightness value;

defining a frame around the one of the plurality of synthetic objects;

calculating a brightness value of the 2D image within an area defined by the predefined frame, and determining the calculated brightness is the same as a brightness of the one of the plurality of synthetic objects; and

inserting the one of the plurality of synthetic objects in the 2D image at a location of the area having a same brightness as the hand image.

7. The method according to claim 1 , wherein the plurality of 2D images come from a same 2D camera.

8. The method according to claim 7 , wherein the 2D camera is a monochrome or an infrared camera.

9. The method according to claim 1 , wherein generating the plurality of synthetic objects comprises generating the plurality of synthetic objects with different physical parameters.

10. The method according to claim 9 , wherein the physical parameters comprise at least one of: a size of a synthetic object, a color of the synthetic object, and accessories arranged on the synthetic object.

11. The method according to claim 1 , wherein the object is a hand.

12. An apparatus, comprising:

a memory storing instructions; and

at least one processor communicatively coupled to the memory, the at least one processor configured to execute the instructions to:

generate a synthetic 3D model of an object with annotation points;

based on the synthetic 3D model of the object, generate a plurality of synthetic objects with different poses and different brightness conditions;

receiving a plurality 2D images with different backgrounds and different brightness conditions;

calculate a brightness value for each of the plurality of 2D images; and

for each of the plurality of 2D images, insert one of the plurality of synthetic objects based on the corresponding brightness value.

13. The apparatus of claim 12 , wherein the at least one processor is configured to execute the instructions to randomly insert the one of the plurality of synthetic objects.

14. The apparatus of claim 12 , wherein the at least one processor is configured to execute the instructions to, for each of the plurality of 2D images, insert the one of the plurality of synthetic objects that presents the same brightness conditions as a corresponding one of the plurality of 2D images.

15. The apparatus of claim 14 , wherein to insert the one of the plurality of synthetic objects, the at least one processor is configured to execute the instructions to:

generate the plurality of 2D images from video acquired using a 2D camera;

model an environment in which the 2D camera operates to acquire the plurality of 2D images with the brightness conditions; and

associate a synthetic object of the plurality of synthetic objects with one of the plurality of 2D images having same brightness conditions.

16. The apparatus of claim 12 , wherein to insert the one of the plurality of synthetic objects, the at least one processor is configured to execute the instructions to:

based on the brightness values, associate a 2D image of the plurality of 2D images with a hand image having a same brightness value;

define a frame around the one of the plurality of synthetic objects;

calculate a brightness value of the 2D image within an area defined by the predefined frame, and determining the calculated brightness is the same as a brightness of the one of the plurality of synthetic objects; and

insert the one of the plurality of synthetic objects in the 2D image at a location of the area having a same brightness as the hand image.

17. The apparatus of claim 12 , wherein the plurality of 2D images come from a same 2D camera.

18. The apparatus of claim 12 , wherein the at least one processor is configured to execute the instructions to generate the plurality of synthetic objects with different physical parameters.

19. The apparatus of claim 12 , wherein the object is a hand.

20. An apparatus comprising:

a memory; and

at least one processor communicatively coupled to the memory, the at least one processor configured to:

receive an image; and

detect an object within the image using a neural network, wherein the neural network is trained based on training data comprising a plurality of 2D images with inserted synthetic objects, wherein the plurality of 2D images with inserted synthetic objects are generated based on a brightness value calculated for each of the plurality of 2D images, wherein, for each of the plurality of 2D images, one of the plurality of synthetic objects are inserted based on the corresponding brightness value.