SARIMEHMETOGLU
About
Contact
Retro-fitting
Yığın (Leftover)
The Alphabet Song
Plotted Leftovers
Knitting
Human Knitting Installation II
CNC Knitting Installation
Human Knitting Installation I
Finger Knitting
Fabric Computing
Robotics
3DGS + Robotics
Verdant Urban
Overprinting
Design for Additive Manufacturing
Cavity Slab
3DGS Slab
Hybrid Computing
Light Leak
Omnifab
Precise Fit
ML as an Instrument
Unrehearsed Compositions
Computational Narratives
Involuntary Shifts
Homage to Erwin Hauer
Sound Installation #1
ZAP!
Machine perception
Counteralgorithms II
Counteralgorithms I
Movement on film
Sequences
Perceived Space
Exhibition Design
Architecture is Art
Space-making
The Interface
Intrastructure
Regenerating a Flux
ULUSKULTUR
Common Ground
MNMTI
2023 - 2024 Fall
Academic + Individual Work
University of Michigan, Ann Arbor - Counteralgorithms
Instructors | Catherine Griffiths
Academic + Individual Work
University of Michigan, Ann Arbor - Counteralgorithms
Instructors | Catherine Griffiths
The old Ulus quarter in Ankara, Turkey, has been a settlement area for many communities with their building genotypes within the urban environment that the city guides. Many communities had settled around the old historical quarter, annotating the name quarter for this particular late 19th and early 20th-century settlement. With the existing history of the area, given its urban context as well, many stories emerge within the existence of the streets, walls, facades, frames, and even disposed goods. Documentation of the area can be understood as less than sufficient. More integrative studies should be conducted to map the stories concerning the area's spatial, architectural, and historical values.
Computer vision can be programmed to understand visual qualities of the area, phenotypes, and the shape grammars of an architectural typology. In addition, with photogrammetry and lidar-based accessible practices applied to a specific settlement, computer vision can map coordinates, materials, meshes, and color values of a given three-dimensional context. Nevertheless, computer vision is less successful than underscoring tangible data in a spatial context, such as the LIDAR output's coordinate values, RGB, location, and point cloud data. In that sense, the project aims to map what computer vision successfully underlines and what it needs to identify further. Both tangible and intangible data are equally effective in understanding genius loci. There is no way to determine which is more important than which—however, computer vision, not being able to complete both intangible and tangible simultaneously, has a side that it chooses: the tangible.
The project stages are mapped in 4 different stages. The first stage includes data collection through photography, underscoring an intangible asset where the human gaze and personal perspective become a part of the frame. In the second stage, LIDAR applications are conducted on the site to get the mesh and material data of certain streets and the significantly mapped home of the father of the neighborhood. This stage is thought to extract tangible values through the existing spatial context, which maps a tangible dataset that is interestingly already processed through computer vision. The third stage includes data processing, where, firstly, datasets mapping the intangible values such as human vision - photographs and videos are processed through computer vision. In the fourth stage, LIDAR datasets extracted with computer vision are annotated through an object detection model through the intensive labor of annotating personal stories on tangible data.
The duality of processing intangible data with computer vision, where already processed tangible data is used as an underlying map for story mapping, creates a new realm of thinking of this already existing space concerning computer vision - what does it detect? Which is more significant to understand the spirit of the spatial context. These two inversely applied data types are located in a 3D space referencing two different project clusters by Forensic Architecture classified by methodology: photogrammetry and 3D modeling. The project aims to understand diverse methods that map the spatial happenings in the office's works and juxtapose them within the same 3D modeling/cinematographic space.