Image processing method and system

1. An image processing method, comprising:

acquiring a sequence of input images containing a target object;

performing multi-resolution fusion on the sequence of input images to generate a single fused image, wherein pixels of the fused image comprise a pixel at a corresponding position of an input image in the sequence of input images, and each pixel of the fused image containing the target object comprises a pixel at a corresponding position of an input image in the sequence of input images in which part of the target object is focused;

applying a 2D fusion algorithm to the sequence of input images to generate an initial fused image; and

receiving a ground truth annotation of the initial fused image to generate an annotated mask map, wherein the annotated mask map indicates whether one or more pixels of the tartlet object in the initial fused image are focused.

2. The method of claim 1 , wherein the acquiring a sequence of input images further comprises:

setting a step size of a camera for acquiring the sequence of input images based on a number of frames of the sequence of input images and a size of the target object in the sequence of input images.

3. The method of claim 1 , wherein the sequence of input images contains indexes, and the performing multi-resolution fusion on the sequence of input images to generate a fused image further comprises:

extracting features of the sequence of input images;

performing multi-resolution fusion on the extracted features, to obtain fused multi-resolution features;

generating a predicted mask map based on the fused multi-resolution features, wherein each pixel of the predicted mask map indicates an index of an input image, and the index indicates an input image from which each pixel of the fused image originates; and

generating the fused image based on the predicted mask map and the sequence of input images.

4. A non-transitory computer readable storage medium including instructions stored thereon which, when executed by at least one processor, cause the at least one processor to perform the image processing method of claim 1 .

5. The method of claim 1 , further comprising:

calculating a loss rate between the predicted mask map and the annotated mask map; and

feeding back the calculated loss rate to a multi-resolution fusion algorithm for performing the multi-resolution fusion.

6. The method of claim 5 , further comprising:

updating the multi-resolution fusion algorithm for performing the multi-resolution fusion based on the loss rate, or the annotated mask map, or a combination of the two.

7. An image processing system, comprising:

an acquirer to acquire a sequence of input images containing a target object;

a fusion circuitry to perform multi-resolution fusion on the sequence of input images to generate a single fused image, wherein pixels of the fused image comprise a pixel at a corresponding position of an input image in the sequence of input images, and each pixel of the fused image containing the target object comprises a pixel at a corresponding position of an input image in the sequence of input images in which part of the target object is focused;

an initial fusion circuitry to apply a 2D fusion algorithm to the sequence of input images to generate an initial fused image; and

an annotation receiving circuitry to receive a ground truth annotation of the initial fused image to generate an annotated mask map, wherein the annotated mask map indicates whether one or more pixels of the target object in the initial fused image are focused.

8. The system of claim 7 , wherein the acquirer is further configured to set a step size of a camera for acquiring the sequence of input images based on a number of frames of the sequence of input images and a size of the target object in the sequence of input images.

9. The system of claim 7 , wherein the sequence of input images contains indexes, and the fusion circuitry further comprises:

an encoder configured to:

extract features of the sequence of input images; and

perform multi-resolution fusion on the extracted features, to obtain fused multi-resolution features; and

a decoder configured to:

generate a predicted mask map based on the fused multi-resolution features, wherein each pixel of the predicted mask map indicates an index of an input image, and the index indicates an input image from which each pixel of the fused image originates.

10. The system of claim 9 , wherein the fusion circuitry is further configured to generate the fused image based on the predicted mask map and the sequence of input images.

11. The system of claim 7 , further comprising:

a loss rate circuitry to:

calculate a loss rate between the predicted mask map and the annotated mask map; and

feed back the calculated loss rate to the fusion circuitry.

12. The system of claim 11 , wherein the fusion circuitry is further configured to update the multi-resolution fusion algorithm for performing the multi-resolution fusion based on the loss rate, or the annotated mask map, or a combination of the two.

13. An image processing system, comprising:

a memory having computer-executable instructions stored thereon; and

a processor coupled to the memory, wherein the computer-executable instructions cause the system to perform the following operations when executed by the processor:

acquiring a sequence of input images containing a target object;

performing multi-resolution fusion on the sequence of input images to generate a single fused image, wherein pixels of the fused image comprise a pixel at a corresponding position of an input image in the sequence of input images, and each pixel of the fused image containing the target object comprises a pixel at a corresponding position of an input image in the sequence of input images in which part of the target object is focused;

applying a 2D fusion algorithm to the sequence of input images to generate an initial fused image; and

receiving a ground truth annotation of the initial fused image to generate an annotated mask map, wherein the annotated mask map indicates whether one or more pixels of the target object in the initial fused image are focused.