suhangpro/mvcnn
Multi-view CNN (MVCNN) for shape recognition
This project helps you classify 3D shapes by analyzing multiple 2D images or "views" of them. It takes inputs like line-drawings, clip art, or rendered 3D models and outputs a general-purpose descriptor that can be used for shape recognition tasks. This is ideal for researchers or engineers working with 3D model databases who need to automatically categorize or search for similar shapes.
388 stars. No commits in the last 6 months.
Use this if you need to reliably identify or categorize 3D shapes based on their visual appearance from various angles, especially when texture information is minimal.
Not ideal if your shapes are primarily defined by detailed textures or if you need to perform direct 3D mesh analysis rather than view-based classification.
Stars
388
Forks
133
Language
MATLAB
License
MIT
Category
Last pushed
Jan 03, 2019
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/suhangpro/mvcnn"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related tools
3DOM-FBK/deep-image-matching
Multiview matching with deep-learning and hand-crafted local features for COLMAP and other SfM...
zouchuhang/LayoutNet
Torch implementation of our CVPR 18 paper: "LayoutNet: Reconstructing the 3D Room Layout from a...
andyzeng/tsdf-fusion-python
Python code to fuse multiple RGB-D images into a TSDF voxel volume.
andyzeng/tsdf-fusion
Fuse multiple depth frames into a TSDF voxel volume.
google/stereo-magnification
Code accompanying the SIGGRAPH 2018 paper "Stereo Magnification: Learning View Synthesis using...