There is one thing that my photography teacher repeated in each of his classes, for two years three times a week: light is the most important thing when photographing, without light you do not have to photograph. I do not intend to refute it, but perhaps your argument is no longer as strong with the advance of technology in better sensors or, as in this case, in artificial intelligence and software.
There are situations in which it is inevitable and we have to take a photograph in low light conditions. Normally, in order to make the elements of the image look as good as possible, two aspects are used: increase the exposure by leaving the shutter open longer or increase the sensitivity of the sensor by ISO. The first option can cause objects to move, the second irremediably adds noise to the image. A third aspect that is increasingly taken into account is the post-processing of images by software.
Several researchers from Intel and the University of Illinois have created a new tool to process poorly exposed images. Like so many other things that arise later, they use automatic learning to recover the photographs. The result, according to the sample images presented by the researchers, is better than what we can achieve by improving the parameters manually in an image processing program.
The researchers took more than five thousand photographs with too short exposure and another five thousand with a good exposure. The AI collected these images and studied them to learn how the images are in darkness and how they should be if they have a good exposure. From this base, the neural network learned, according to the researchers, to reconstruct images taken almost in complete darkness.
If normally we simply increase the exposure, what is done is to illuminate pixels with lighter shades. This causes the “white dots” characteristic of any images taken at night. Artificial intelligence seems to be applying a series of techniques in the image, in addition to increasing exposure, also “color” the areas with the most appropriate color possible, which is the differential factor. In contrast, it blurs and loses small details since it is not able to correctly define the limits of each area.
This new tool is in its early stages of development, it has yet to improve its results but they are already astonishing if one takes into account that the images are practically dark. Software processing has more and more weight when it comes to taking a picture and the clearest example of this is in smartphones. Rivalizan with a DSLR camera in many occasions and not by hardware, but to improve the images by software later.
For more stuff visit our site techverses.com and discover what you want.
I am a 24 years old boy looking forward to becoming the top blogger in the world. I work efficiently because my writing is influenced by the broader experience of the world, and I know how to work well with people and deadlines.