Overview

An assistant to a visually challenged person for boarding a train can be of great service. With advance image recognition techniques powered by AI’s deep learning aspect and with the mobile camera, the problem is solved with only an application on the mobile. The features of the app should include:

• ● Near Real-time image processing
• ● Digital Indicators for visually challenged people to comprehend
• ● Available with just a soft upgrade on the current handsets
• ● Easy to use
• ● Offline – work while the network is away

NIIT Technologies’ Deep Learning implementation with the latest features from the mobile library is empowering the client to provide these services.




BUSINESS IMPERATIVES

In the world of billions of people, there are few, who are unable to utilize the current rail infrastructure at the fullest due to their challenged vision. A digital assistant at their service with no additional financial / social burden this challenge can be easily overcome

Apart from serving the needy, this can be a great opportunity for the service provider to:

• ● Enhance their brand image
• ● Create quick media bytes
• ● Looked upon as modern AI warehouse



ARCHITECTURE

NIIT Technologies uses the easily available tech components with some advance deep learning techniques to provide this application.



Coreml is latest available library from Apple to easily integrate the machine learning models into an ios app. The benefits are:

• ● Optimized for on-device performance
• ● Minimized memory footprint and power consumption
• ● Running on mobile device
• ● Ensures the privacy of user data
• ● Guarantees the functionality of app even without network availability

Keras is an open-source neural network software library capable of running TensorFlow for machine learning. It is used to detect and decipher patterns and correlations, analogous to the learning and reasoning which humans use.

Amazon Elastic Compute Cloud (Amazon EC2) provides a secure, resizable compute capacity in the cloud for utilizing the GPUs on need basis. It makes easier for the developer to configure the hardware / software capacity quickly.