---
title: "Object Detection & Recognition"
description: "Detect faces, recognize embeddings, and decode QR codes and barcodes using Iris's built-in detection pipelines."
keywords: ["object detection", "face detection", "face recognition", "QR code", "barcode", "embedding", "ArUco"]
---
# Object Detection & Recognition
Iris provides built-in pipelines for object detection, face recognition, barcode/QR reading, and ArUco marker detection.
## Face Detection & Recognition
### Detection
Detects human faces, bounding boxes, confidence scores, and facial landmark coordinates (left/right eyes, nose, left/right mouth).
```rust
let detector = FaceDetector::<Backend>::default();
let faces = detector.detect(&image)?;
for face in faces {
println!("Face found with confidence: {}", face.confidence);
println!("Bounding box: {:?}", face.bbox);
}
```
### Recognition (Embeddings)
Extracts unique 512-dimension face embedding vectors and computes similarity metrics.
```rust
let recognizer = FaceRecognizer::<Backend>::default();
// Extract embeddings
let emb1 = recognizer.extract_embedding(&face_img1)?;
let emb2 = recognizer.extract_embedding(&face_img2)?;
// Compute similarity score (cosine distance)
let score = recognizer.compute_similarity(&emb1, &emb2)?;
println!("Face similarity: {}", score);
```
## QR & Barcode Detection
### QR Code Detector
```rust
let qr_detector = QrDetector::new();
let qr_codes = qr_detector.detect_and_decode(&image)?;
for qr in qr_codes {
println!("QR decoded: '{}'", qr.payload);
println!("Corners: {:?}", qr.corners);
}
```
### Barcode Detector
```rust
let barcode_detector = BarcodeDetector::new();
let barcodes = barcode_detector.detect_and_decode(&image)?;
for bc in barcodes {
println!("Barcode payload: '{}' (format: {})", bc.payload, bc.format);
}
```
## ArUco Marker Detection
```rust
let aruco = ArucoDetector::new(ArucoDict::Dict6x6_250);
let markers = aruco.detect_markers(&image)?;
for marker in markers {
println!("Marker ID: {}", marker.id);
}
```
## Object Detection (General)
```rust
let detector = ObjectDetector::<Backend>::default();
let detections: Vec<Detection> = detector.detect(&image)?;
for det in detections {
println!("Detected class {} at {:?}", det.class_id, det.bbox);
}
```