Deep Learning for Travel: Gifting Sight to Those Without It

Carol, 36, works as a client-servicing executive at a travel agency. She is different from her colleagues as she is visually challenged. Every morning, she walks from her residence in Fremont, CA to Thornton Avenue to catch a bus to the Bay Area Rapid Transport (BART) station. It is a daily challenge for her to recognise the door to the train when she lands at the BART station, as she has to rely on others next to her for the same. It is an equally big challenge for her when she gets off the train at San Francisco. This is a real problem faced every day by visually impaired passengers globally And, there are millions like her around the world with special needs who struggle to lead their lives hassle-free.

According to WHO, there are over 300 million visually impaired people around the world. The American Foundation for the Blind estimates there are at least 1.3 million such people who are employed in the US alone! While figures for the rest of the world are not established, various estimates indicate that at least 200 million visually challenged individuals globally are either working or educating themselves to get into the workforce. This consequently implies that organizations must gear themselves up fast to service this massive customer base with a differing set of requirements.

Inclusive Customer Experience is the Way Forward

According to various studies, differently-abled people comprise anywhere between two to three percent of the world’s workforce. Visually challenged persons make a sizable portion of this talent pool. If organizations are focusing on providing superior customer experiences, then their effort should extend to embrace even the differently-abled customers.

With advances in technology, it is possible for organizations to build tools and techniques that can make lives easier for their differently-abled customers. New technologies like Artificial Intelligence (AI) and Deep Learning help you build tools that will make life easier for your customers with disabilities. Let us consider how the implementation of Deep Learning can make life easier for visually impaired customers.

The Anatomy of Deep Learning-based Train Solution

With Deep Learning capabilities, a Convolutional Neural Network Model for image recognition and scene classification can be developed. The video of the train is converted into a set of images, and Deep Learning algorithms can classify these images as door / no door and sends the output to the passenger’s iPhone. The phone then vibrates to tell the passenger that the train door is in front of him.

Images can be trained over AWS and a mobile device is used only for inference. The model works in an offline environment, which makes it convenient to use even without a network, too.

Deep Learning in the Service of Visually Impaired

It’s Monday. Carol walks out to the BART station in Fremont, CA. However, things are going to be different today. She uses a mobile app that leverages advanced image recognition techniques aided by Machine Learning capabilities of AI that connect to her mobile phone camera. She no longer needs to depend on others to detect the entrance or the exit door to the train. Her app provides her near real-time image processing and digital indicators that she can comprehend despite her visual disability. She can use it offline too. And the app is available on the handset of her choice. She does not have to go binging on a new device.

This is a fine illustration of how organizations can enhance end-user experience by building applications, which are emotionally empathetic and seek to offer an inclusive brand experience even to customers with special needs. It is important that you work with a technology provider with a track record of working on cutting-edge technology and delivering applications that make a difference to customers, regardless of their impairments.

Making Lives Better in the Future

While technologies like AI and Deep Learning have the potential to radically change the way we lead our lives, it is still a work in progress. Look at the kind of work that is going into the domain of autonomous cars. Researchers are trying to breathe intelligence into mechanical objects using AI and other assistive technologies. While AI and Deep Learning have the potential to overcome problems and make lives a lot easier, there are many limitations to be overcome. Solving the problems and prevailing over the technological limitations will eventually lead us in the direction of empowering not merely the machines, but also people with special needs. It is an evolution in which we all need to participate.