Method and system for generating an entirely well-focused image of a large three-dimensional scene

1. A method of generating an entirely well-focused image of a three-dimensional scene, the method comprising the steps of:

a) learning a prediction model including at least a focal depth probability density function (PDF), h(k), for all depth values k, from one or more historical tiles of the scene;

b) predicting the possible focal surfaces in a subsequent tile of the scene by applying the prediction model;

c) for each value of k, examining h(k) such that

if h(k) is below a first threshold, no image is acquired at the depth k′ for said one tile; and

if h(k) is above or equal to a first threshold, one or more images are acquired in a depth range around said value of k for said one tile; and

d) processing the acquired images to generate a pixel focus map for said subsequent tile.

2. The method as claimed in claim 1 , wherein step c) comprises examining h(k) such that

if h(k) is below the first threshold, no image is acquired at the depth k′ for said one tile; and

if h(k) is above or equal to the first threshold and below a second threshold, one or more images are acquired in the depth range around said value of k for said one tile using a first sampling rate; and

if h(k) is above or equal to the second threshold, one or more images are acquired in the depth range around said value of k for said one tile using a second sampling rate higher than the first sampling rate.

3. The method as claimed in claim 1 , further comprising updating the prediction model before steps a) to d) are applied to a next neighboring tile.

4. The method as claimed in claim 1 , comprising the steps of

i) for a first tile, acquiring images at equally spaced values of k and processing the acquired images to find a pixel focus map for said first tile;

ii) building the PDF based on said pixel focus map for said first tile;

iii) applying steps a) to d) for n consecutive neighboring tiles; and

for a (n+1)th tile, return to step i) treating the (n+1)th tile as the first tile.

5. The method as claimed in claim 1 , wherein the PDF is a pre-learned model and/or a user defined model.

6. The method as claimed in claim 1 , wherein the prediction model comprises a structure component and a probabilistic component.

7. The method as claimed in claim 1 , wherein the acquiring of images comprises capturing images or reading stored images.

8. The method as claimed in claim 1 , wherein the method is applied to microscopy or photography.

9. The method as claimed in claim 1 , wherein the learning of the prediction model comprises using spatial contextual information.

10. A system for generating an entirely well-focused image of a three-dimensional scene, the system comprising:

a learning unit for learning a prediction model including at least a focal depth probability density function (PDF), h(k), for all depth values k, from one or more historical tiles of the scene;

a prediction unit for predicting the possible focal surfaces in a subsequent tile of the scene by applying the prediction model;

a processing unit for, for each value of k, examining h(k) such that

if h(k) is below a first threshold, no image is acquired at the depth k′ for said one tile; and

if h(k) is above or equal to a first threshold, one or more images are acquired in a depth range around said value of k for said one tile; and

d) processing the acquired images to generate a pixel focus map for said subsequent tile.

11. A computer readable data medium having stored thereon a computer code means for instructing a computer to execute a method of generating an entirely well-focused image of a three-dimensional scene, the method comprising the steps of:

a) learning a prediction model including at least a focal depth probability density function (PDF), h(k), for all depth values k, from one or more historical tiles of the scene;

b) predicting the possible focal surfaces in a subsequent tile of the scene by applying the prediction model;

c) for each value of k, examining h(k) such that

if h(k) is below a first threshold, no image is acquired at the depth k′ for said one tile; and

if h(k) is above or equal to a first threshold, one or more images are acquired in a depth range around said value of k for said one tile; and

d) processing the acquired images to generate a pixel focus map for said subsequent tile.