A Google Glass Based Real-Time Scene Analysis for the Visually Impaired

Blind and Outwardly Impeded Individuals (BVIP) are probably going to encounter challenges with assignments that include scene acknowledgment.Wearable innovation plays had a huge impact in exploring and assessing frameworks created for and with the BVIP people group.This paper presents a framework in light of Google Glass intended to help BVIP with scene acknowledgment undertakings, subsequently involving it as a visual colleague.The camera implanted in the shrewd glasses is utilized to catch the picture of the environmental elements, which is examined utilizing the Custom Vision Application Programming Connection point (Vision Programming interface) from Purplish Blue Mental Administrations by Microsoft.The result of the Vision Programming interface is switched over completely to discourse, which is heard by the BVIP client wearing the Google Glass.A dataset of 5000 recently clarified pictures is made to work on the presentation of the scene depiction task in Indian situations.The Vision Programming interface is prepared and tried on this dataset, expanding the mean Typical Accuracy (Guide) score from 63% to 84%, with an IoU > 0.5.The general reaction season of the proposed application was estimated to be under 1 second, accordingly giving precise outcomes continuously.A Likert scale investigation was performed with the assistance of the BVIP instructors and understudies at the ''Roman and Catherine Lobo School for the Outwardly Impeded'' at Mangalore, Karnataka, India.From their reaction, it very well may be reasoned that the application assists the BVIP with bettering perceiving their general climate progressively, demonstrating the gadget's power as an expected colleague for the BVIP.