Dog-Yang

Dog-Yang / ConInfer

Public

CVPR 2026 (Findings)-ConInfer: Context-Aware Inference for Training-Free Open-Vocabulary Remote Sensing Segmentation

19
0
100% credibility
Found Apr 07, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

ConInfer provides a training-free method using vision-language models to perform open-vocabulary semantic segmentation on remote sensing imagery for tasks like building, road, and water extraction.

How It Works

1
🔍 Discover ConInfer

You find this tool while searching for ways to automatically identify features like buildings or roads in satellite photos.

2
📥 Gather your images

Collect your aerial or satellite images and matching maps if you have them, or use sample datasets provided.

3
⚙️ Set up the analyzer

Follow simple steps to prepare your computer so it can process the images.

4
🎯 Choose what to find

Pick the type of feature you want to spot, like roads, buildings, or water bodies.

5
▶️ Start the analysis

Click to run and let it scan your images, watching progress as it works.

✅ View your map

See colorful overlays highlighting exactly where features are, with scores showing confidence.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 19 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is ConInfer?

ConInfer delivers training-free open-vocabulary segmentation for remote sensing images, letting you query arbitrary categories like "roads" or "buildings" in satellite or aerial photos using vision-language models. It processes large scenes with Python scripts, outputting masks and metrics to spreadsheets via a simple multi-GPU eval command. Backed by the CVPR 2026 accepted paper (Findings), it tackles inconsistent patch predictions in real-world overhead imagery.

Why is it gaining traction?

Unlike pixel-independent baselines, ConInfer models spatial and semantic context across scenes, boosting accuracy by 2.8% on segmentation and 6.13% on extraction tasks versus SegEarth-OV. Developers grab it for zero-shot results on 15+ datasets without training, amid buzz around CVPR 2026 papers github and CVPR 2026 reddit discussions. The github cvpr 2026 template fits neatly into research pipelines.

Who should use this?

Remote sensing devs segmenting unlabeled aerial data for urban planning or disaster response. CVPR 2026 hopefuls prototyping OVRSS ahead of the deadline. GIS teams needing quick masks for roads, water, or buildings without custom models.

Verdict

Solid pick for training-free remote sensing seg if you're in CVPR 2026 workshops or chasing cvpr papers github—beats baselines handily. But 19 stars and 1.0% credibility score signal early maturity; setup needs dataset prep tweaks, docs are sparse.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.