svilupp / OmniParserIconDetectors.jl

MIT License

1 stars 0 forks source link

feat: Initial implementation #1

Closed devin-ai-integration[bot] closed 4 days ago

devin-ai-integration[bot] commented 1 week ago

OmniParserIconDetectors.jl Implementation

This PR implements a Julia wrapper for the OmniParser icon detection model.

Features

Automatic model download via DataDeps.jl
Image preprocessing to 640px on longest side
Direct ONNXRuntime.jl integration (no Python dependencies)
Coordinate translation to original image space
Visualization with confidence indicators
Comprehensive test suite

Detection Parameters

Minimum confidence threshold: 15%
Minimum detection size: 15x15 pixels
Maximum aspect ratio: 5.0
NMS IoU threshold: 0.2

Dependencies

Requires specific ONNXRunTime.jl version:

Pkg.add(url="https://github.com/svilupp/ONNXRunTime.jl#mac120")

Testing

All tests are passing with proper error handling and validation.

Link to Devin run: https://preview.devin.ai/devin/f9269b1272d7491ab2e3a5e71c594058

If you have any feedback, you can leave comments in the PR and I'll address them in the app!