Google's Veo 3.1 can turn separate images into a single video
SMRTR summary
Animators once spent countless hours crafting motion frame by painstaking frame. Now Google's latest AI can weave together random images into flowing video clips in seconds.
Google DeepMind just released Veo 3.1, a video-generating model that works like a visual blender, combining separate images into what the company calls unified "smoothies" of motion. Feed it a photo of a woman's face, some clothes, and an ornate room, and it produces a short clip of her strolling through that space wearing those garments.
The AI gets wonderfully surreal when given seemingly incompatible images. Google demonstrated this by combining a Christmas tree behind sliding doors with swirling psychedelic colors. The result: doors opening to release a flood of multicolored ornament-sized balls, like a festive reimagining of The Shining's elevator scene.
The model can also work with just two images, automatically filling the gaps between a first and final frame. Users can now extend video clips or add and remove visual elements from existing footage, dramatically shrinking production timelines that once required teams of animators.
SMRTR provides this summary for quick context. The original article belongs to ZDNet.
Read the original article