How is Lumen5 powered by A.I.?
We're building Lumen5 as a video creation platform with the goal of offering the world's simplest video making experience, and our primary approach to this is to have the workflow assisted by AI. The idea is that the user would simply enter a link to a blog post or news article and Lumen5 will automatically turn the content into an engaging video. Lumen5 is designed for brands and businesses to produce video content at scale through automation.
The first bit of AI you'll encounter is when you enter a link to a blog post or news article where our natural language processing algorithms will strive to help you summarize the article. As you know, a full article contains too much text for a short and engaging video, hence summarization is necessary. Currently, the system will detect and divide up various sections within the article and offer them as suggestions to you. There's also a Smart Summarization feature which will help you summarize long run-on sentences. In the future, we hope to improve this algorithm further so that little to no human decisions are required, and all you get is a perfectly summarized storyboard for your video.
The next bit of AI you'll encounter is our computer vision algorithms that automatically matches each of your scenes with corresponding background media such as photos, graphics, and videos. For example, if you entered an article about basketball, then the system will automatically seek out basketball-related media files and add them to your video. This area of intelligence will improve over time, allowing us to provide more and more accurate media matches and suggestions.
Many video details are also controlled by AI as well, such as the duration of each scene being dynamically determined based on the number of words in that scene. Lumen5 will automatically detect the language you're creating your video in, and then calibrate the timing based on average reading speed in that language. Computer vision is used to detect the focus of an image, such as human faces, and the system automatically positions your text to avoid covering the main subject of the background media. The system will also detect important phrases and keywords, then automatically apply a coloured highlight to bring attention to those words.
Last but not least, perhaps one of the most valuable component of Lumen5 is the built-in media library consisting of millions copyright-free photos and videos along with thousands of audio tracks. The toughest part to creating video content is securing all the assets you need to produce a great video. Lumen5 does it all for you and you can search through the database with keywords, then drag-and-drop them into your video along with an automatically generated credit scene at the end to ensure all attributions are properly in place.
The result is the ability to create engaging social videos within minutes rather than more traditional workflows which could take hours, or days!