Object aware local tone mapping

1. A system comprising:

an image sensor configured to capture an image; and

a processing apparatus configured to:

access the image from the image sensor;

detect an object area on the accessed image;

classify the detected object area into a first type object area or a second type object area;

obtain a local tone mapping for the second type object area of the image;

apply a bilateral filter to the second type object area of the image to obtain a low-frequency component image portion and a high-frequency component image portion;

obtain, for the second type object area, a second enhanced image portion based on a weighted sum of the low-frequency component image portion and the high-frequency component image portion, wherein the high-frequency component image portion is weighted more than the low-frequency component image portion;

obtain, for the second type object area, an enhanced image portion based on the local tone mapping; and

store, display, or transmit an output image based on the enhanced image portion and the first type object area.

2. The system of claim 1 , wherein the processing apparatus is further configured to:

obtain, for the second type object area, a third enhanced image portion based on the second enhanced image portion and a tone mapping.

3. The system of claim 2 , wherein the processing apparatus is further configured to:

obtain, for the second type object area, a perceptual domain image portion based on the third enhanced image portion and a gamma curve that models human perception of contrast.

4. The system of claim 3 , wherein the processing apparatus is further configured to:

obtain, for the second type object area, a low-frequency component perceptual domain image portion and a high-frequency component perceptual domain image portion as components of the perceptual domain image portion.

5. The system of claim 4 , wherein the processing apparatus is further configured to:

obtain, for the second type object area, a fourth enhanced image portion based on a weighted sum of the low-frequency component perceptual domain image portion and the high-frequency component perceptual domain image portion, wherein the high-frequency component perceptual domain image portion is weighted more than the low-frequency component perceptual domain image portion.

6. The system of claim 1 , wherein the first type object area is classified as a human face.

7. The system of claim 1 , wherein the first type object area is classified as a blue sky.

8. The system of claim 1 , wherein the processing apparatus is configured to enhance local contrast on the second type object area of the image.

9. The system of claim 1 , wherein the processing apparatus is further configured to perform perceptual tone mapping, lens distortion correction, electronic rolling shutter correction, stitching with parallax correction, blending, electronic image stabilization, or output projection.

10. The system of claim 9 , wherein the processing apparatus is configured to perform the perceptual tone mapping with a gamma curve.

11. A method comprising:

obtaining an image from an image sensor;

detecting an object area on the image;

classifying the detected object area into a first type object area or a second type object area;

applying a filter to the second type object area to obtain a low-frequency component image portion and a high-frequency component image portion;

determining a first enhanced image portion for the second type object area based on a weighted sum of the low-frequency component image portion and the high-frequency component image portion, where the high-frequency component image portion is weighted more than the low-frequency component image portion;

determining a second enhanced image portion for the second type object area based on the first enhanced image portion and a tone mapping; and

storing, displaying, or transmitting an output image based on the second enhanced image portion and the first type object area.

12. The method of claim 11 , in which determining the second enhanced image portion for the second type object area based on the first enhanced image portion and a tone mapping comprises:

applying the tone mapping to the low-frequency component image portion to obtain gains for respective image portions; and

applying the gains for respective image portions to corresponding image portions of the first enhanced image portion.

13. The method of claim 11 , further comprising:

determining a perceptual domain image for the second type object area based on the second enhanced image portion and a gamma curve that models human perception of contrast;

determining, for the second type object area, a low-frequency component perceptual domain image portion and a high-frequency component perceptual domain image portion as components of the perceptual domain image;

determining, for the second type object area, a third enhanced image portion based on a weighted sum of the low-frequency component perceptual domain image portion and the high-frequency component perceptual domain image portion, where the high-frequency component perceptual domain image portion is weighted more than the low-frequency component perceptual domain image portion; and

wherein the output image is based on the third enhanced image portion and the first type object area.

14. The method of claim 13 , in which determining the low-frequency component perceptual domain image portion comprises applying a transformation, based on the gamma curve, to a result of applying the tone mapping to the low-frequency component image portion.

15. The method of claim 13 , in which determining the output image comprises:

determining gains for respective image portions based on the third enhanced image portion and the gamma curve; and

applying the gains for respective image portions to corresponding image portions of the image.

16. An image capture device comprising:

an image sensor configured to capture an image; and

a processing apparatus configured to:

access the image from the image sensor;

detect an object area on the accessed image;

classify the detected object area into a first type object area or a second type object area;

apply a filter to the second type object area to obtain a low-frequency component image portion and a high-frequency component image portion;

apply a non-linear mapping to the low-frequency component image portion to obtain gains for respective image portions;

apply the gains for respective image portions to corresponding image portions of the second type object area to obtain an enhanced image portion; and

store, display, or transmit an output image based on the enhanced image and the first type object area.

17. The image capture device of claim 16 , in which the processing apparatus is configured to:

determine the non-linear mapping based on a histogram analysis of image portions of the low-frequency component image portion.

18. The image capture device of claim 16 , in which the processing apparatus is configured to:

determine, for the second type object area, a perceptual domain image portion based on the enhanced image and a gamma curve that models human perception of contrast;

determine, for the second type object area, a low-frequency component perceptual domain image portion and a high-frequency component perceptual domain image portion as components of the perceptual domain image;

determine, for the second type object area, an enhanced perceptual domain image based on a weighted sum of the low-frequency component perceptual domain image portion and the high-frequency component perceptual domain image portion, where the high-frequency component perceptual domain image portion is weighted more than the low-frequency component perceptual domain image portion; and

wherein the output image is based on the enhanced perceptual domain image and the first type object area.

19. The image capture device of claim 18 , in which the processing apparatus is configured to determine the low-frequency component perceptual domain image, the processing apparatus is further configured to apply a transformation, based on the gamma curve, to a result of applying the gains for respective image portions to the low-frequency component image portion.

20. The image capture device of claim 18 , in which the processing apparatus is configured to determine the output image, the processing apparatus is further configured to:

determine gains for respective image portions based on the enhanced perceptual domain image and the gamma curve; and

apply the gains for respective image portions to corresponding image portions of the image.