Detection Head

A detection head acts as the final decision-making layer in an object detection neural network architecture. While the earlier layers of the model are responsible for understanding the shapes, textures, and features within an image, the detection head is the specific component that interprets this information to predict exactly what objects are present and where they are located. It transforms the abstract, high-level data produced by the feature extractor into actionable results, typically outputting a set of bounding boxes enclosing identified objects along with their corresponding class labels and confidence scores.

Distinguere la testa dalla colonna vertebrale e dal collo

To fully grasp the function of a detection head, it is helpful to visualize modern detectors as being composed of three primary stages, each serving a distinct purpose in the computer vision (CV) pipeline:

• Backbone: This is the initial part of the network, often a Convolutional Neural Network (CNN) like ResNet or CSPNet. It processes the raw input image to create feature maps that represent visual patterns.
• Neck: Sitting between the backbone and the head, the neck refines and combines features from different scales. Architectures like the Feature Pyramid Network (FPN) ensure the model can detect objects of varying sizes by aggregating context.
• Head: The final component that consumes the refined features from the neck. It performs the actual task of classification (what is it?) and regression (where is it?).

Evoluzione: basata su ancoraggi vs. senza ancoraggi

The design of detection heads has evolved significantly to improve speed and accuracy, particularly with the transition from traditional methods to modern real-time inference models.

• Teste basate su ancoraggi: i rilevatori di oggetti tradizionali a stadio singolo si basavano su riquadri di ancoraggiopredefiniti, ovvero forme di riferimento fisse di varie dimensioni. La testa prevedeva di quanto allungare o spostare questi ancoraggi per adattarli all'oggetto. Questo approccio è descritto in dettaglio nella ricerca fondamentale su Faster R-CNN.
• Anchor-Free Heads: State-of-the-art models, including the latest YOLO26, utilize anchor-free detectors. These heads predict object centers and dimensions directly from the pixels in the feature maps, eliminating the need for manual anchor tuning. This simplifies the architecture and enhances the model's ability to generalize to novel object shapes, a technique often associated with Fully Convolutional One-Stage Object Detection (FCOS).

Applicazioni nel mondo reale

The precision of the detection head is critical for deploying artificial intelligence (AI) in safety-critical and industrial environments. Users can easily annotate data and train these specialized heads using the Ultralytics Platform.

• Guida autonoma: nell' intelligenza artificiale per il settore automobilistico, la testa di rilevamento ha il compito di distinguere in tempo reale tra pedoni, semafori e altri veicoli. Una testa altamente ottimizzata garantisce che la latenza di inferenza rimanga sufficientemente bassa da consentire al veicolo di reagire istantaneamente.
• Diagnostica medica: nell' analisi delle immagini mediche, le testine di rilevamento sono ottimizzate per individuare anomalie come i tumori nelle scansioni MRI. Il ramo di regressione deve essere estremamente accurato per delineare i confini esatti di una lesione, assistendo i medici nelle soluzioni sanitarie.

Esempio di codice

The following example demonstrates how to load a YOLO26 model and inspect the output of its detection head. When inference runs, the head processes the image and returns the final boxes containing coordinates and class IDs.

from ultralytics import YOLO

# Load the YOLO26n model (nano version)
model = YOLO("yolo26n.pt")

# Run inference on an image to utilize the detection head
results = model("https://ultralytics.com/images/bus.jpg")

# The detection head outputs are stored in results[0].boxes
for box in results[0].boxes:
    # Print the bounding box coordinates and the predicted class
    print(f"Class: {int(box.cls)}, Coordinates: {box.xywh.numpy()}")

This interaction highlights how the detection head translates complex neural network activations into readable data that developers can use for downstream tasks like object tracking or counting.