Researchers at Intel have used machine learning AI to make GTA V count insanely realistic .

" Photorealism has been the limit finish of figurer graphics for half a 100 , " the teamwrite in their paper . " Nevertheless , a face at even the most advanced real - time games will quickly let out that photorealism has not been achieved . An unspeakable conflict in the show of simulation and reality continue . "

They go down out to make a system - construction on previous attempts using machine learning - that could achieve this form of photorealism .

During the translate operation , game engines produce intermediate buffers known as " G buffers " . Essentially , this provides elaborated data point on what will terminate up on screen , such as geometry , kindling and textures . The main innovation of this algorithm over others is that the squad used this intermediate data to train their   algorithmic program and improve the output signal double .

The squad also used photographs accumulate around Germany by cars with built in cameras ( exchangeable to Google Maps ) calledCityscapesas training for their neural connection to raise the image output signal .

former attempts at enhancing biz graphics as it is rendered have relied on computer labeling of objects . As a answer , it ’s often unable to be render at usual plot speeds , or can bring forth additional unwanted picture , such as the trees seen in the skyline demonstrated in the video recording below .

" or else of judge to synthesize images , our approach path enhances already render images , integrates scene entropy to produce geometrically and semantically consistent images , and does not need any annotation of tangible information , " the team excuse in their report .

The results of their work are quite stupefying , provide footage that looks pretty close to photorealism , or at least photorealistic enough to make you feel uncomfortable about go on a received GTA V kill fling .

The blending is better , the Hill a caboodle gullible , and roads a lot smoother to front at . At time - though not always - it in short looks like you are ram around a real city .

The squad believe that their system could be improved by integrate it more deeply into game railway locomotive , " increase efficiency and perchance further further the point of realism " . More astonishing double produced by their system of rules can be viewedon their GitHub Sir Frederick Handley Page .