How can we use wide-FOV cameras for reconstruction?. We propose self-calibration Gaussian Splatting that jointly optimizes camera parameters, lens distortion, and 3D Gaussian representations to directly reconstruct from a set of wide-angle captures. page:

RT @GoogleDeepMind: What if you could not only watch a generated video, but explore it too? 🌐. Genie 3 is our groundbreaking world model th….

Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.

RT @PMinervini: "in 2025 we will have flying cars" 😂😂😂

RT @gene_ch0u: We've released all code and models for FlashDepth! It produces depth maps from a 2k, streaming video in real-time. This was….

RT @gradeeterna: Training directly on the circular fisheyes from 360 cameras like @insta360 X5 and @KandaoVR Qoocam 3 Ultra using @NVIDIAAI….

RT @Almorgand: "UFM: A Simple Path towards Unified Dense.Correspondence with Flow". TLDR: End-to-end trained transformer model; directly re….

RT @zhenjun_zhao: We'll be resenting our #CVPR2025 Best Paper Award Candidate GlobustVP today!. 🎤 Oral Session 4C: 3D Computer Vision.🗓️ Sa….

RT @omooretweets: AI ASMR is blowing up . This account was started three days ago with just 11 videos…but has racked up 82k followers 👇 htt….

RT @iam_NCJ: What if a Transformer could render?.Not text → image. But mesh → image — with global illumination. No rasterizers. No ray-tra….

RT @HavenFeng: When it comes to recovering a dynamic world, many researchers focus on extracting geometry and camera parameters—via SfM or….

RT @junyi42: Introducing St4RTrack!🖖. Simultaneous 4D Reconstruction and Tracking in the world coordinate feed-forwardly, just by changing….

RT @xichen_pan: We find training unified multimodal understanding and generation models is so easy, you do not need to tune MLLMs at all. M….

RT @LFC:

RT @minchoi: This is wild. This is NOT CGI or sped up 🤯. China's EngineAI humanoid robot just nailed the "Axe Gang" dance from Kung Fu Hus….

RT @UnitreeRobotics: World's First Side-Flipping Humanoid Robot: Unitree G1 😊.One year after Unitree H1 (1.8m) pioneered the first standing….

😮.

RT @janusch_patas: FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Views. Contributions:.- We propo….

The work is done with amazing team: @wenqi_xian, @stevenygd, Leonidas Guibas, @GordonWetzstein, Steve Marschner, @debfx.

We also enable fast camera parameter optimization alongside reconstruction. It is integrated into the original CUDA rasterizer for efficient optimization.

Here, I show comparisons of reconstruction from undistorted images vs our method. Our pipeline achieve more complete FOV reconstruction.

We propose using a hybrid field to model lens distortion while applying cubemap resampling to preserve uniform pixel density.