# Object detection Get 2D bounding-box detections from a camera feed using ML or color heuristics, track objects across frames, and act on or alert on results. > Source: https://docs.viam.com/vision/object-detection/ Detect 2D objects in camera frames: get bounding boxes with labels and confidence scores, follow objects across consecutive frames, and wire results into machine behavior or notifications.
Detect objects

Retrieve 2D bounding-box detections from a vision service, filter by confidence and class, and run detections in a loop.

Detect by color

Configure the color_detector vision service to find regions of a specific hue. No ML model, no training data, runs anywhere.

Tune detection quality

Match a detection quality symptom to the mlmodel vision service attribute that fixes it. Covers bounding box order, input normalization, per-label thresholds, and tensor remapping.

Track objects

Track detected objects across consecutive video frames with persistent IDs using the object-tracker module.

Act on detections

Build a module that uses vision service results to control machine behavior.

Alert on detections

Send email, webhook, or push notification alerts when your vision service detects specific objects or classifications.