In recent times, text analytics is proving insufficient to get a complete perspective about brands or products discussed on social media and other online locations. According to Michalis Michael, CEO, DigitalMR, importantly, images or videos that come with the text should also be taken into account to get the complete story.
Increasingly, we see posts with little to no text. For instance, the posts on Instagram are just one or more images. Therefore, text analytics alone might prove insufficient. One way to analyze these images and videos is to leverage a deep learning approach. Deep learning—a form of machine learning—can be used to detect the theme and context of an image and describe it using text. In addition, the audio component in the videos can be converted into text by leveraging automatic speech-to-text technology.
How does the deep learning approach help in such instances? With the ability of deep learning to turn pixels into text, it is possible to know about the image, cluster it under its respective discussion bucket, understand the consumers’ opinion about the brand or product category, and identify usage occasions or discover exclusive insights on the competition.
Click here to read the article.
Please give your feedback on this article or share a similar story for publishing by clicking here.
