Algorithm’s analysis of an city

How environment is perceived by an algorithm ? From found footage of Strasbourg city walk tour I ask algorithm to decompose the image in what it see via depth prediction, image segmentation, object detection and pose estimation. I also added datamoshing to the scene.

Datamosh is about adding movement information of the next shot. Thus overlapping/overwriting information.  There is this kind of overwriting on Machine learning training — where the neurons weight are changing when learning. There is also this kind of overwriting when finetuning aka rewrite learning on pretrained models.

With these kind of image, I need to pursue with idea outside of the plasticity of the render but also what does it tell ?