Oxford Town Centre dataset

Oxford Town Centre is a dataset created by researchers at Oxford University for the research and development of pedestrian activity and facial recognition systems. 

The data was collected in 2009 from a public safety CCTV camera in the middle of Oxford and captured the movements of approximately 2,000 people.

Operator: Amazon; Disney; OSRAM; Huawei
Developer: University of Oxford
Country: UK
Sector: Govt - municipal; Research/academia; Technology
Purpose: Improve pedestrian detection
Technology: Dataset; Computer vision; Facial recognition; Pattern recognition
Issue: Privacy; Dual/multi-use; Surveillance
Transparency: Governance; Marketing; Privacy

Risks and harms 🛑

The Oxford Town Centre dataset has proved popular, having been used in over 60 verified research projects including commercial research by Amazon, Disney, OSRAM, and Huawei; and academic research in China, Israel, Russia, Singapore, the US, and Germany among dozens more. It has been downloaded over 700 times on Kaggle. 

As Exposing.ai activist Adam Harvey pointed out in September 2019, the dataset is 'unique in that it uses footage from a public surveillance camera that would otherwise be designated for public safety. Yet research citations show that the footage has been used in dozens of research projects with no connection to public safety, and that most of the research took place outside the UK. Pedestrians appearing in the video act normally and unrehearsed suggesting that the images were captured without consent.'

The dataset has also been used to develop public and office social distancing algorithms, notably by US company Landing AI, which posted a demonstration video featuring the Oxford Town Centre dataset. The video was deleted on YouTube after it was highlighted by Exposing.ai, though Landing AI CEO Andrew Ng's tweet promoting the demo remains online.

The dataset was removed from the University of Oxford's website in June 2020. However, it remains easily and freely accessible on Kaggle and other data sharing communities.

Derivatives, applications 🈸

Investigations, assessments, audits 🧐

Page info
Type: Dataset
Published: February 2023