Crate od_opencv 

Object detection utilities for YOLO-based neural networks.

This crate provides wrappers for running YOLO models.

Backends

• opencv-backend (default): Uses OpenCV DNN for inference. Supports Darknet and ONNX models.
• ort-backend: Uses ONNX Runtime for inference. Pure Rust, no OpenCV required.

Supported Models

• Traditional YOLO (v3, v4, v7) - Darknet format, opencv-backend only
• Ultralytics YOLO (v8, v9, v11) - ONNX format, both backends

Re-exports

pub use bbox::BBox;

pub use image_buffer::ChannelOrder;

pub use image_buffer::ImageBuffer;

pub use model_trait::ObjectDetector;

pub use model_factory::Model;

pub use backend_ort::ModelUltralyticsOrt;

pub use backend_ort::OrtModelError;

Modules

backend_ort

ONNX Runtime backend for object detection.

bbox

Bounding box type for object detection results.

image_buffer

Image buffer type for storing image data.

model_factory

Model factory for creating object detection models.

model_trait

Backend-agnostic object detection trait.

postprocess

Post-processing utilities for object detection.

preprocessing

Preprocessing utilities for object detection.