This slide shows how lighting conditions affect the appearance of the image. Different settings will produce different image noise. This explains why we would like to do the data augmentations (adding white noise and etc) when dealing with image-like input data.
micahtyong
This photo showcases an interesting example where part of the image exhibits low lighting conditions (e.g., buildings in the background, especially in the upper left hand corner), and some parts of the image exhibits very bright lighting conditions (e.g., area on the right where the fireworks occur). I wonder if a similar technique to mipmapping is used to apply variable exposure lengths to different parts of an image.
This slide shows how lighting conditions affect the appearance of the image. Different settings will produce different image noise. This explains why we would like to do the data augmentations (adding white noise and etc) when dealing with image-like input data.
This photo showcases an interesting example where part of the image exhibits low lighting conditions (e.g., buildings in the background, especially in the upper left hand corner), and some parts of the image exhibits very bright lighting conditions (e.g., area on the right where the fireworks occur). I wonder if a similar technique to mipmapping is used to apply variable exposure lengths to different parts of an image.