[ad_1]
Honor 10
One of the newest flags of AI on the Bulgarian market is illumination to create beautiful images. The outside light reaches the camera lens in an uncontrollable way. However, with the relevant knowledge of lenses, filters, the environment (film or sensor) and lighting, the photographer can tame light and photograph stunning images by changing camera settings and by adding the most appropriate filters
they are often narrowly specialized in disciplines such as portraiture, product, landscape, architecture, fashion and many other types of photography. It is because mastering an aspect of photography takes many years, if not a life. help
Segmentation of the Semantic Image
Segmentation of the semantic image is a technique designed to know what "what" is part of every pixel in the image (path, people, trees, etc.).
Today, the segmentation of images by artificial intelligence chambers mainly covers two scenarios of use. First, the effect Boke – distinguish objects from the background by the blur of the background. Second, by recognizing the "type of scene" (nature, flowers, food, etc.) to put the camera in special automatic mode or preset automatic mode, the Honor smartphone camera can do more that and use the semantic image segmentation to create a hierarchy of information that can be used for much finer camera parameters, and this in real time with a neutral processing unit (NPU).
Imagine that you want to create group joins. in front of an incredible landscape and you would like everyone to be in the center if the background is not you can use the boke effect that will enhance the photo.You can not create a general rule for photos without knowing the context of the image.If the effect of boke is applied, the camera must ensure that everyone in the group will also be in the center of attention, even if someone One is a little behind the AI.
Revealing facial beauty
Another example of how Honor's camera uses artificial intelligence is to illuminate the face in all conditions. If you make the sunset in the background, even the artificial intelligence of the previous generation camera can go into Sunset mode. But while this guarantees good color settings to accentuate the sunset, the subject's face may look too bright. If the camera switches to Selfie mode, the sunset may not be optimally taken. With the semantic segmentation of the image, the artificial intelligence distinguishes the two objects in the image and applies different settings to each one of them to get the best picture.
The Honor camera also supports 3D face recognition as well as other distinctive facial features such as nose, cheekbones, etc. In this way, the camera can ensure that the contours of the face will be clear and at the same time preserve the natural beauty of the skin. This is a significant advance over previous versions of Beautify, which are less intelligent and tend to blur too much, making them unnatural.
Perhaps one of the most significant achievements of mobile portrait photography is the use of portrait lighting that mimics professional lighting as in a photo studio. Without this 3D camera, some researchers have described this extras as extremely difficult to achieve, but the Honor NPU makes portrait lighting possible and effective. or Google uses techniques such as image segmentation in the manufacture of devices. However, the segmentation is only the foundation of Honor's innovation. The ability to recognize different objects in the image is one, but Honor engineers use this information to perform a complex camera configuration in real time using its own methodology.
AI photography is the the most difficult and energy-intensive computing task which Honor phones use a separate NPU processor to handle these tasks instead of the common processor. Internal tests have shown that the Honor 10 neural processing unit (NPU) is 25 times faster and 50 times more powerful than some processors. Excellent speed and efficiency make it possible to obtain more accurate data with excellent performance and energy efficiency. Flagman Honor 10 is on sale in Bulgaria since June of this year. It can be purchased from the Technopolis network for 899 and 999 BGN.
Computer photography and artificial intelligence not only contribute to shooting but also to the processing of photographs: retouching, editing, animation. For example, imagine how a group shot can be the result of several shots – AI can create the final shot using the best picture of everyone – with a beautiful smile and eyes open
In fact, AI is already a valuable aid for professional designers in the selection of objects in photo editing applications. The selection has become a very difficult and tedious task that can now be performed with one click AI. The selection is an important step in the photorealistic treatment that allows the user to apply filters only to certain parts of the photo.
Professional photographers can take incredible pictures and enjoy the technology for their very large number of sequences. These photos must be reviewed, selected and modified. In this process, AI can save you a lot of time. Artificial intelligence could allow the computer to view and select (or delete) images thousands of times faster than humans. Once selected, AI can edit them on a near-professional level. Such innovations are now in the testing phase and show impressive results
Copyright © CROSS Agency Ltd.
When using the content of the CROSS information agency
the reference is mandatory.
[ad_2]
Source link
