BubbleID: Image Segmentation Model for Boiling Analysis Released on HuggingFace
On September 14, 2026, the model UARK-NED3/BubbleID was published on HuggingFace with the image-segmentation task. It is a research package for analyzing liquid–vapor boundaries in high-speed video sequences of bubble boiling.
What it means
The model is lightweight: the largest file — seg_model.pth — takes up 0.3 GB, and the repository holds about 0.4 GB of weights in pth format in total. It runs via PyTorch. The license is dryad-record-10.5061-dryad.ksn02v7gx, tied to a Dryad record. The developer has not published any quantized builds.
BubbleID combines segmentation, tracking and classification for analyzing bubbles and interfaces. The source test sequences were captured at 3000 fps in steady state and at 150 fps in transient state. For installation, launch and training materials, the author refers to the GitHub repository.
Fresh news notes about 3D and video tools that can be run on your own hardware are collected in the CPU3D news feed.How the method works. The diagram is drawn based on this news note.
H3 Cinematic Multishot Coverage Released on HuggingFace for Image-to-VideoOn August 24, 2026, the ethanfel/H3_Cinematic_Multishot_Coverage model was published on HuggingFace for the image-to-video task. The code is released…
UniMoCa unifies human motion and camera control in a single visual representationResearchers have introduced UniMoCa, a method that translates both human motion and camera trajectory into a shared visual representation called Motio…
DiffVC-ONE: One-Step Video Compression with Video Diffusion TransformerarXiv published DiffVC-ONE, a generative video compression framework based on a one-step Video Diffusion Transformer. The authors describe three compo…
VA-Judger: Reward Model for Joint Video and Audio GenerationResearchers introduced VA-Judger, a reward model for post-training of joint video and audio generation systems, along with the VAPref-10K dataset of 1…