CPC G06T 5/92 (2024.01) [G06N 3/045 (2023.01); G06T 5/50 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/20208 (2013.01); G06T 2207/20212 (2013.01)] | 14 Claims |
1. A method for generating an HDR image comprising:
generating, based on a brightness value of each pixel included in an original image, a first image including an area having the brightness value greater than or equal to a specific value in s the original image and a second image including a remaining area except for the area having the brightness value greater than or equal to the specific value in the original image; and
generating a high dynamic range (HDR) in which a dynamic range of the original image is extended from the first image and the second image using a pre-trained model including a first neural network, a second neural network, and a third neural network,
wherein the first neural network is pre-trained to output a third image in which a range of the brightness value of the first image is extended in a direction in which the brightness value is increased based on the first image,
the second neural network is pre-trained to output a fourth image in which a range of the brightness value of the second image is extended in a direction in which the brightness value is decreased based on the second image, and
the third neural network is pre-trained to generate the HDR image based on the third image and the fourth image,
wherein the generating of the first image and the second image includes:
generating a mask based on the brightness value of each pixel,
generating the first image by masking the original image using the mask, and
generating the second image by inverting brightness of the first image, and
wherein in the generating of the mask, the mask is generated using Equation 1 below,
where P(i,j) is a brightness value corresponding to (i, j) pixel of the original image, and threshold is a threshold value determined based on the brightness value of each pixel.
|