OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

PolygonRNN++

PolygonRNN++ is a pretrained computer-vision model and inference code that predicts polygon boundaries around image regions in crops, supporting instance-segmentation dataset annotation.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

PolygonRNN++ — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use PolygonRNN++ when you need locally generated polygon predictions to support image-crop annotation for segmentation datasets.

Who it is for

  • Researchers and teams creating instance-segmentation datasets.
  • Developers who want to run polygon prediction locally on image crops.

Documented uses

  • Generate polygon outlines as initial instance annotations for image crops.
  • Run the provided inference demo and examine its output files.
  • Follow a notebook walkthrough for running the model on sample crops.
Before you choose

Requirements & limits

The repository is public and licensed under GPL-3.0.

  • The documented workflow uses TensorFlow and requires downloading 448 MB of pretrained model files and graphs.
  • GPU use is recommended; the repository says CPU inference works but is slow.
  • The pretrained models were trained on Cityscapes data.
  • The README directs users seeking training and tool code to a separate official PyTorch reimplementation.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up and run the sample inference demo

Prepare the repository's documented local environment and run its demo inference script.

Before you start

  • A Linux environment with Python and virtualenv available.
  • GPU and tensorflow-gpu are recommended; the README says CPU execution is possible but slow.
  • 448 MB for downloading the pretrained models and graphs.

Steps

  1. Clone the repository and change into its directory.

  2. Create and activate a virtual environment, then install dependencies from requirements.txt.

  3. Run ./models/download_and_unpack.sh to download and unpack the pretrained models and graphs.

  4. Run ./src/demo_inference.sh.

Expected outcome

The README says the demo should produce results in the output/ folder.

Guide limitations

  • CPU execution is described as slow.
  • The downloaded models were trained on the Cityscapes dataset.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
PolygonRNN++Generate polygon outlines as initial instance annotations for image crops.Linux, Self-hostedThe repository is public and licensed under GPL-3.0.
Opencv ZooFind an OpenCV DNN-oriented model for a visual task such as face detection, object tracking or image segmentation.Windows, Linux, macOSThe repository is public and lists an Apache-2.0 license. The README says the project is now hosted on Hugging Face for models and online demos.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
7 November 2026
Original website title
fidler-lab/polyrnn-pp · Official homepage title · retained in its source language
How we review sourcesSuggest a correction