Hugging Face's logo

Spaces:
huzey
/
ncut-pytorch
Running on Zero

App Files Community
1
ncut-pytorch / README.md
huzey's picture
huzey
update config
226b4fb
preview code
|
raw
history blame
1.46 kB
metadata
title: ncut-pytorch
emoji: ✂️
colorFrom: yellow
colorTo: pink
sdk: gradio
sdk_version: 4.42.0
app_file: app.py
pinned: false
license: apache-2.0

Documentation https://ncut-pytorch.readthedocs.io/

NCUT: Nyström Normalized Cut

Normalized Cut, aka. spectral clustering, is a graphical method to analyze data grouping in the affinity eigenvector space. It has been widely used for unsupervised segmentation in the 2000s.

Nyström Normalized Cut, is a new approximation algorithm developed for large-scale graph cuts, a large-graph of million nodes can be processed in under 10s (cpu) or 2s (gpu).

Gallery

TODO

Installation

PyPI install, our package is based on PyTorch, presuming you already have PyTorch installed

pip install ncut-pytorch

Install PyTorch if you haven't

pip install torch

Why NCUT

Normalized cut offers two advantages:

  1. soft-cluster assignments as eigenvectors

  2. hierarchical clustering by varying the number of eigenvectors

Please see NCUT and t-SNE/UMAP for a full comparison.

paper in prep, Yang 2024

AlignedCut: Visual Concepts Discovery on Brain-Guided Universal Feature Space, Huzheng Yang, James Gee*, Jianbo Shi*, 2024

Normalized Cuts and Image Segmentation, Jianbo Shi and Jitendra Malik, 2000