ZERO DEPTH
MACHINE VISION - POSITION RECOMPOSED
TEAM : Sanhita S Vartak, Shabnam Moravveji, Hsiao-Chiao Peng
Zero Depth
Starting from studding Ledoux’s design principles and architectural elements following by Comme de garcon’s ways of assembling the pieces, as building blocks of a proxy project, we took advantage of 4 ledoux’s elements to investigate on different generative technique’s in representing objects.
In our first step, we tried to compressed our 4 solid object’s orthographical captures from top view together and building super images with zero sense of depth. our post processing technique was basically playing with different pieces from each grid of nine with separated channels in blue and red in order to extract various information from original pieces and achieve a series of super images that could generate our profiles in representing new object by multiple projection.
Robot house workflow - is important to create a repeatable process in order to understand machine vision in each project
Considering dimensionalizing as a transformation ,we studied to allocate a particular points of view to machines and manipulate it’s observation from the real object; By deliberately diminishing the input information as a technique, we could become able to receive different readings from our actual objects via digital tools and post processing softwares, therefore in second step , via a series of transformational processes including overlapping each objects capture frames ,we compressed all observational information that machine can absorb in a series of capturing frames through particular path in order to create super images that can eventually help us in recreation of series of objects that contain multiple layers of observed information by machine. In this stage, we used various Tanique’s and software’s that can provide parallel results by considering machine’s multiple point of view, first one of of them was combining the super images of our first approach with the second one and creating new series of profiles for multiple projection, the next one was using varying layers of super images as a surface in 3d coat software that gave us another series of 3d objects that contains multiple layers of information.
By comparing these two results, we can apply different visible qualities to compare and select for the next steps. So far we compressed all the captured information in the super images and used them as guiding profiles in creation of our new objects, for the next step, we are considering decompression of our new objects to extract different existing layers of information. By cutting our new meshes derived from 3dcoat and extracting sections of information, projecting them on our parallel profiles we can generate a series of new objects.
