CloudSoftwareHardware
Software Stack
DepthAI
OAK Apps
AI Models
Troubleshooting

ON THIS PAGE

  • OAK Examples
  • Overview
  • Generic Example
  • Custom Examples

OAK Examples

Overview

OAK Examples is a collection of inference examples that can be used as a starting point or source of inspiration for applications of your own. The examples utilize the concepts introduced in the Inference section, and run real-time AI inference on the OAK devices. We suggest to first explore the Generic example and continue with numerous Custom examples we have prepared for various ML tasks.

Generic Example

The Generic Example runs a simple inference pipeline of a single-head model predicting on a single-image input. For example, you can run YOLOv6 object detection on your camera input as:
Python
1python3 main.py --model luxonis/yolov6-nano:r2-coco-512x288

Custom Examples

We also offer a diverse collection of custom examples. that demonstrate the use of pipelines with multiple models (which may have multiple heads) and/or custom output processing. These examples span a wide range of ML tasks, including:
  • Classification: Assigns labels, such as identifying objects in images (e.g., cat vs. dog);
  • Object Detection: Identifies and locates multiple objects within an image;
  • Face Detection: Analyzes face features, expressions, and behaviors;
  • 3D Detection: Extends object detection into 3D space, estimating the object position, size, and orientation;
  • Keypoint Detection: Detects specific points of interest within an object (e.g. face landmarks);
  • Pose Estimation: Analyzes human or animal bodies, predicting joint positions and posture;
  • Segmentation: Labels each pixel according to its category (e.g., objects vs. background);
  • Object Tracking: Tracks movement of individual objects across a sequence of images in the same scene;
  • Counting: Counts the number of objects within an image (e.g., people in a crowd);
  • Optical Character Recognition: Extracts text from images, converting it into machine-readable text;
  • Reidentification: Matches and recognizes individual objects across images of different scenes;
  • Depth Estimation: Predicts the distance of objects from the camera, creating a depth map of a scene;
  • Line Detection: Identifies lines or edges in images (e.g., lane detection in autonomous driving);
  • Image-to-Image Translation: Transforms images from one format to another (e.g., enhancing resolution);
  • Feature Detection: Identifies key features or points in an image, used for tasks like image matching;
  • Speech Recognition: Converts spoken language into text by analyzing audio signals.
We showcase some of the examples below but make sure to check the full collection on GitHub.