Google is now improving the functionality of its Veo 3 video generation model with the inclusion of image-to-video generation. On Thursday, the Mountain View-based tech giant announced that eligible users will now be able to add images as input and generate videos based on them using Veo 3. Notably, this feature is currently available via Gemini’s web interface, and the mobile apps do not support it. Currently, Google is releasing this capability in select countries, but it has not specified the names of these countries.
Veo 3 Can Now Generate Videos Using Image Input
In a blog post, the tech giant announced that Veo 3 will now support image inputs while generating videos. This means users can create artificial intelligence (AI) videos of any reference image they like. There is some confusion regarding its availability, as Google has only said that it is being released in “select countries.” Gadgets 360 staff members were able to confirm that the feature has not yet rolled out in India.
Veo 3’s image-to-video generation feature is not available in India
Those who get access to this new feature can go to the Gemini website, select the “Video” chip located underneath the text box, and an option to add photos should now be visible. Here, users can add the reference image and write a prompt detailing how they would like to animate the video, and the Veo 3 model takes care of the rest.
Google also shared a demo video highlighting the range of complexity one can add to their prompts. In the short video, Veo 3 took an image of a cardboard box and generated videos of a hamster cooking a meal inside it, a man jumping inside the box, and a lift coming out of it.
Notably, all videos generated using Gemini will include a visible watermark to highlight that these are not real videos, as well as an invisible watermark added using Google’s SynthID technology. The latter cannot be cropped out, edited, or removed via any other means. These only show up in special software designed to read the watermark.
Veo 3 is currently available in 154 countries and territories globally. The video generation model can be accessed via the Google AI Pro and Google AI Ultra subscriptions. The Ultra subscription, which costs $200 (roughly Rs. 17,200) a month, is only available in the US.
