CAAD Praxis: Cats in Zurich, dogs in Berlin, trees in New York

Cats in Zurich, dogs in Berlin, trees in New York: Object localization in Satellite Images to generate Semantic Maps for Urban Design

052-0629-18 G  | Elective course | CAAD Praxis

Today, satellite and aerial imagery are fundamentally available for nearly any point on the planet. Besides this, during the last years, considerable efforts have been made to develop various methods for the detection of different types of objects in urban aerial data.

In this course, we will dig into the abundance of urban aerial imagery and create personal semantic maps of any city, by learning how to talk to cities and articulate questions around our own interests. We will introduce you to Convolutional Neural Network (CNN), show you how to localise specific objects on satellite images and generate semantic maps of cities. Cats in Zurich, dogs in Berlin, trees in New York or horses in Paris, what is at stake here is your own question. Make cities meaningful by projecting on them what actually matters to you!

participation max. 20, min.10 motivated students
dates Mondays, 15:00–17:00
introduction Monday, 24 September 2018
place Chair for CAAD, HIB E 16
course tutor  Karla Saldaña Ochoa, Zifeng Guo

Lecture 01, Introduction

Lecture 02, The interplay of Big Data and Machine Intelligence

Lecture 03, Data Driven Modeling for Urban Applications

Lecture 04, Map and Model

Lecture 05, Introduction to the experiment

lecture 06, Collecting Data

Lecture 07, Label Data

Lecture 08, Train you own model

Lecture 09, Rendering your model