Image rendering method and apparatus

1. An image rendering method for rendering a pixel at a viewpoint, comprising the steps of:

(a) for a first element of a virtual scene, having a predetermined surface at a position within that scene, and for each respective one of a plurality of contributing components of the image:

(i) providing the position and a direction based on the viewpoint to a respective machine learning system previously trained to predict a respective factor for the respective contributing component that, when combined with a respective distribution function that characterises an interaction of light with the predetermined surface for the respective contributing component, generates a respective pixel value for the respective contributing component corresponding to the first element of the virtual scene as illuminated at the position for the respective contributing component; and

(ii) combining the respective predicted factor from the respective machine learning system with the respective distribution function to generate the respective pixel value for the respective contributing component corresponding to the illuminated first element of the virtual scene at the position;

(b) combining the respective pixel values generated for each respective one of the contributing components of the image to create a final combined pixel value; and

(c) incorporating the final combined pixel value into a rendered image for display;

wherein the respective machine learning systems were previously trained with a training set based on images comprising multiple lighting conditions.

2. The image rendering method according to claim 1 , in which: the multiple lighting conditions comprise one or more changes to lighting position.

3. The image rendering method according to claim 2 , in which: a change in lighting position is a function of one or more of:

i. notional time within the virtual scene; and

ii. a motion path of a movable virtual light source.

4. The image rendering method according to claim 1 , in which: the multiple lighting conditions comprise one or more changes to lighting direction.

5. The image rendering method according to claim 4 , in which: a change in lighting direction is a function of one or more of:

i. a change in axial direction of a light source; and

ii. a change in angular spread of light.

6. The image rendering method according to claim 1 , in which the multiple lighting conditions comprise one or more changes to one or more of:

i. lighting colour;

ii. lighting brightness; and

iii. lighting diffusion.

7. The image rendering method according to claim 1 , in which

the machine learning system is a neural network;

an input to a first portion of the neural network comprises the position of the predetermined surface; and

an input to a second portion of the neural network comprises the output of the first portion and the direction.

8. The image rendering method according to claim 7 , in which if inputs to all or part of the neural network represent more than one property of light, then an at least partially connected additional input layer is provided to the neural network.

9. The image rendering method according to claim 1 , in which

the machine learning system is a neural network; and

an input to the neural network comprises data representative of one or more of:

i. a lighting position or offset;

ii. a notional time within the virtual scene;

iii. axial direction of a light source;

iv. angular spread of light;

V. colour or colour temperature;

vi. brightness; and

vii. diffusion.

10. The image rendering method according to claim 1 , comprising

selecting at least a first trained machine learning model from among a plurality of machine learning models, the machine learning model having been trained, with a training set based on images comprising multiple lighting conditions, to generate data contributing to a render of at least a part of an image;

wherein the at least first trained machine learning model has an architecture based learning capability that is responsive to at least a first aspect of a virtual environment for which it is trained to generate the data; and

using the at least first trained machine learning model to generate data contributing to a render of at least a part of an image.

11. The image rendering method according to claim 1 , comprising the steps of

using at least two respective machine learning models trained on different respective lighting conditions;

generating a pixel value corresponding to the first element of the virtual scene as illuminated at the surface position using the respective machine learning models; and

combining the generated pixel values as the output of the machine learning system.

12. The image rendering method according to claim 1 , comprising

using a machine learning model trained with a training set generated by the steps of:

generating a plurality of candidate viewpoints of a scene;

culling candidate viewpoints according to a probability that depends upon a response of the surface of the scene to light at a surface position in the scene corresponding to the viewpoint; and

generating training images at the remaining viewpoints.

13. A non-transitory, computer readable storage medium containing a computer program comprising computer executable instructions, which when executed by a computer system, cause the computer system to perform an image rendering method for rendering a pixel at a viewpoint by carrying out actions, comprising:

(a) for a first element of a virtual scene, having a predetermined surface at a position within that scene, and for each respective one of a plurality of contributing components of the image:

(i) providing the position and a direction based on the viewpoint to a respective machine learning system previously trained to predict a respective factor for the respective contributing component that, when combined with a respective distribution function that characterises an interaction of light with the predetermined surface for the respective contributing component, generates a respective pixel value for the respective contributing component corresponding to the first element of the virtual scene as illuminated at the position for the respective contributing component; and

(ii) combining the respective predicted factor from the respective machine learning system with the respective distribution function to generate the respective pixel value for the respective contributing component corresponding to the illuminated first element of the virtual scene at the position;

(b) combining the respective pixel values generated for each respective one of the contributing components of the image to create a final combined pixel value; and

(c) incorporating the final combined pixel value into a rendered image for display;

wherein the respective machine learning systems were previously trained with a training set based on images comprising multiple lighting conditions.

14. An entertainment device, comprising

a graphics processing unit configured to render a pixel at a viewpoint within an image of a virtual scene comprising a first element having a predetermined surface at a position within that scene;

a machine learning processor configured to provide, for each respective one of a plurality of contributing components of the image, the position and a direction based on the viewpoint to a respective machine learning system previously trained to predict a respective factor for the respective contributing component that, when combined with a respective distribution function that characterises an interaction of light with the predetermined surface for the respective contributing component, generates a respective pixel value for the respective contributing component corresponding to the first element of the virtual scene as illuminated at the position for the respective contributing component;

the graphics processing unit being configured to combine, for each respective one of the plurality of contributing components of the image, the respective predicted factor from the respective machine learning system with the distribution function to generate the respective pixel value for the respective contributing component corresponding to the illuminated first element of the virtual scene at the position;

the graphics processing unit being configured to combine the respective pixel values generated for each respective one of the contributing components of the image to create a final combined pixel value; and

the graphics processing unit being configured to incorporate the final combined pixel value into a rendered image for display;

wherein the respective machine learning systems were previously trained with a training set based on images comprising multiple lighting conditions.