[08] Perception
[08.i] Counteralgorithms II
The project investigates the historic Ulus quarter in Ankara as a layered urban settlement shaped by multiple communities, architectural genotypes, and lived stories that remain under-documented. It argues that understanding the area’s spatial, architectural, and historical value requires integrating both tangible data (geometry, materials, coordinates) and intangible data (stories, human perspective, memory).
The study explores how computer vision interprets this environment, emphasizing that while it excels at extracting tangible information—especially through LiDAR, photogrammetry, point clouds, and RGB data—it struggles to capture intangible qualities related to genius loci. Rather than privileging one over the other, the project frames their coexistence and tension as productive.
The methodology unfolds in four stages:
-
Photography-based data collection, foregrounding human gaze and subjective experience as intangible input.
-
LiDAR scanning of selected streets and a key neighborhood house to extract tangible spatial and material data.
-
Computer vision processing of photographic and video data to analyze intangible visual patterns.
-
Annotation of LiDAR datasets, where personal and communal stories are manually embedded into tangible spatial data using object detection.
By inversely applying computer vision to intangible and tangible datasets within a shared 3D modeling environment, the project questions what computer vision can and cannot perceive, and how spatial meaning emerges from this divide. It situates this inquiry alongside methodologies used by Forensic Architecture—particularly photogrammetry and 3D modeling—aiming to juxtapose multiple modes of spatial narration within a single cinematic 3D space.