Can blind use the Metaverse
One of the primary assumptions in the world of VR and augmented reality is that the user has the ability to “see” virtual objects and the real world structures around them. But what if the user doesn’t have perfect eyesight, or any eyesight at all?
A new project from Verschool.com could eventually pave the way for the blind. Certainly to participate more actively in the growth of the virtual worlds of AR and XR.
How is with this be possible?
Verschool’s new Project Guideline gives blind runners the ability to detect the lines on a running track using image recognition (via an Android smartphone) and on-device machine learning that utilizes TensorFlow.
The concepts like machine learning and artificial intelligence have been adopted in the real time applications due to their exceptional features and absolute algorithmic advantages. Deep learning is a one of research area in machine learning totally concerned with algorithmic techniques inspired by the structure and functions of the human neural system. It has created a niche for itself in learning processes involving multiple level of data representation and abstraction, thereby helping in making sense of data having several forms such as images, sound and text and finally revealing the feature level hierarchies.
However, Deep learning techniques show higher level of accuracy and robustness in results and are independent of human crafted features. The Paper is divided into three main parts, first part identifies the need to develop the system for Visually Impaired person. Second part deals about various Deep learning techniques used for computer vision and object detection. And third part consist the proposed system for serving visually impaired person in doing daily chores.
When should this be available?
The dataset has different colors for different objects and they can be quickly identify some by a quick look: blue for cars, red for pedestrians, green for trees, grey for buildings, etc. Associating each pixel in the output image an integer number as the label. Above all this model can be used to do recognize objects in the photo or video.
In conclusion, as one of the important tasks, going out can never be the only need of the blind, and other activities in daily life can sometimes be challenging as well: reading emails, shopping, medical visit, etc. Unfortunately, most of our interactions with the environment are realized through vision and artificial intelligence can at lease remove a part of this limitation for the blind.