Qualcomm lately announce its latest flagship chopine for smartphones , theSnapdragon 8 Gen 2 .
One of the biggest update here was the young Cognitive ISP , a smart trope signal processor that delivers up to 200 - megapixel images , more level-headed expression detection from the Always - Sensing photographic camera and all the benefits of real - time semantic segmentation .
But , what is semantic segmentation and what does it mean for your phone ’s photographic camera ?
[ squirrel asin = B09NRXBD9H ]
What is semantic segmentation?
Semantic segmentation is an image engineering that enable a camera to recognise individual vista within a frame , such as faces , tomentum , clothes , backgrounds and more . Those aspects can then be optimised on an individual basis , entail the camera will modify the colour , look , asperity and amount of noise in the image on a case - by - case basis .
Think of semantic sectionalisation like Photoshop layers within your earpiece camera , allowing different parts of an prototype to be adjust individually rather than the photographic camera editing the image as a whole .
Real - sentence semantic sectionalisation is on the nose what it sounds like – the same as the above but everything befall as you use the photographic camera and not after you click the pic or land up register the TV .
There are loads of instances in which real - fourth dimension semantic segmentation can improve your fluid picture taking .
The tech can be used to smooth peel , remove reflections from glasses , sharpen hair and fur , work more low-spirited to the sky and better the legibility of school text on screen and in Book , to name a few real - world employment .
you may see semantic segmentation in action in the video above .
How accurate is it?
We had the chance to test material - time semantic segmentation at Qualcomm ’s Snapdragon Summit in Hawaii this year and come up it did an okay job of separating skin from hair , apparel and the scope behind us , but it was n’t perfect .
In the demonstration , the camera was set to turn everything it take up to be skin green to give us an estimate of how precise it is .
The photographic camera did a unspoiled job of highlight all the skin in the figure of speech , aside from a modest section peeking out from behind the glass . However , it struggled a mess with likewise color elements , erroneously highlight the beige clothes and orange iPhone as skin because their tone of voice meet too nearly with the facial expression and arms in the image .
In another demonstration , we test the skin smoothing lineament and found it to be very elusive and born , resulting in something a bit more convincing than some of the more overzealous smoothing effects found on TikTok , Instagram and some OEM ’s Android camera software .
Ultimately , it ’s down to each OEM to tune the feature article enabled by the Cognitive ISP the means they like , VP of product management for camera Judd Heape secernate us at the Summit , mean we ’ll require to wait for the first batch of 8 Gen 2 phones to go far to see how real - clip semantic division will improve images with real - world use .
