Generating Live Captions from a Camera Feed

As humans, we see a million objects around us every day, in different scenarios. For humans, describing a scene is usually a trivial task: something we do without even taking a noticeable amount of time to think. But it is a huge task for machines to comprehend the elements and scenarios presented to it in visual media such as images or videos. However, for several applications of artificial intelligence (AI), it is useful to have the capability of comprehending such images in the computer system. For example, it would be of immense help to visually impaired people if we could devise machines that could translate their surroundings into audio in real time. Also, there has been a constant effort from researchers to generate captions for images and videos in real time, so as to improve the accessibility of content presented on websites...