Researchers have developed a technique called Gaussian Splatting that transforms videos into 3D representations, aiding in video tracking, depth prediction, and editing. The method involves rasterizing depth and compressing splats efficiently.
Trimming the Fat: Efficient Compression of 3D Gaussian Splats through Pruning https://t.co/LCf9Jwbnw2 https://t.co/edge9prl9z
On Scaling Up 3D Gaussian Splatting Training https://t.co/Wl6tgegKu2 Project: https://t.co/umHDH9mSQx Code (Apache 2.0): https://t.co/WjhXyvXl0X => Parallelize training over multiple GPUs. Make sure to checkout the project page, it is awesome! Method ⬇️ 1 | 2 https://t.co/julArdgUsT
RaDe-GS: Rasterizing Depth in Gaussian Splatting. https://t.co/rqANY8d82F
What a timely thread on Gaussian splats! Must see for every roboticist 🤖 https://t.co/EJdBJTxhEB
Great thread on Gaussian Splatting for robotics https://t.co/AM7Sgvrlaz
Splatter a Video can turn a video into a 3D Gaussian representation! This allows for enhanced video tracking, depth prediction, motion and appearance editing, and stereoscopic video generation. https://t.co/lbXBe2hkJ3 https://t.co/SrahRVf99B