SenseCity Africa – new technologies to collect urban data
SenseCity Africa explores the potential of telecommunication technologies to address various urban challenges across the African continent. The initial project phase involved gathering information through the participation of the general public. This consisted of sharing photos of different African cities taken on-site by participants, who became “human sensors” able to take an active part in representing the nuances of their cities and cultures.
Around 1,500 geo-referenced photographs and 500 tags were collected. These visual and textual contributions provided an overview of urban and peri-urban landscapes covering 28 African countries. The results were presented at the Seoul Biennale of Architecture and Urbanism in 2021, then in 2022 at the 3rd edition of World Cities Day in Yaoundé, Cameroon, and at EPFL for the 20th anniversary of the School of Architecture, Civil and Environmental Engineering (ENAC).
This dataset forms the foundation for the ongoing phases that are now unfolding. We are currently deeply engaged in comprehensive studies to assess state-of-the-art computer vision (CV) models. Our focus lies in evaluating these models to identify and subsequently address biases and limitations inherent in CV algorithms when employed within African contexts. Therefore, SenseCity Africa serves as a precursor to the creation of a comprehensive training dataset, and more accurate (and inclusive) CV models.
General information
Team: Ximena Salgado Uribe, Vitor Pessoa Colombo, Jérôme Chenal
Contact: Ximena Salgado Uribe
Funding: Self-funded (EPFL-CEAT)