Going to the bar with Artifical Intelligence

Olivia Brookhouse    26 December, 2019

The average adult spends 2 months of their life queuing at bars and pubs for drinks and at Christmas these queues can seem even longer. Whilst some of us are patient, it is estimated that the average adult is only happy to wait up to 8 minutes to place their order, putting pressure on establishments to optimise their services. For bartenders, the challenge to pour pints quickly means they are often unable to identify who is next in line to be served, meaning certain customers are left waiting far longer. But a new Artificial Intelligece technology developed by DataSparQ is about to make going out a stress free experience, eliminating queue jumpers and making getting numerous drinks a lot easier.

The product developed by DataSparQ, which uses facial recognition technology, identifies those who arrived first and places customers into a digital queue, making it harder for queue jumpers to skip ahead. The digital queue is projected onto a screen above the bar so that customers can identify where they are in the queue and therefore how long it will take to order a drink. The list is also displayed behind the bar for the bartenders. Being too small or shy to push your way to the front are now irrelevant thanks to the intelligent system developed by DataSparQ.

The technology developed by DataSparQ is simpler than those installed in security systems as it able to detect faces within the live video stream but it is unable to recognize the identity of the customers as it does not have access to this data. Recognizing a face is a face and not an orange is a long way away from recognizing a face is a face and therefore who it belongs to. This is the difference between facial detection and facial recognition.

Facial recognition technology uses Machine Learning, specifically Deep Learning to identify or verify a person from a digital image or video. Deep Learning algorithms consist of many layers of neurons tied together, forming a virtual neural network. Each face has 80 distinguishable characteristics that make it unique, including the distance between your nose and brows and the shape of your mouth which the algorthm will turn into a mathematical code. The analysis of these individual feautures is turned into code, formed by mathematical formulas similar to that of a fingerprint, but far more complex.

Facial recognition technology is often the subject to controversy for being too intrusive. In many cases the technology has not been developed with diverse enough data sources and therefore the algorithm is unable to recognize all ages or races. The EU has promised to take a strong stance on the issue to adhere to their General Data Protection Regulation (GDPR), to protect citizens from the growing use of surveillance technology in public spaces, especially security firms collecting and storing unlimited amounts of data.

However, the product developed by DataSparQ promises to delete all data after 24 hours and therefore does not have the ability to identify any of the faces. EU legislation to ensure the correct use of facial recognition will mean those companies who are developing innovative and useful applications of the technology such as DataSparQ are not viewed in the same way as those abusing it.

To stay up to date with LUCA, visit our Webpage, subscribe to LUCA Data Speaks and follow us on TwitterLinkedIn YouTube.

Leave a Reply

Your email address will not be published.