Table of Contents

Device and platform support

Thanks to cloud-side computation, the cloud image recognition feature has minimal hardware and platform requirements, making it ideal for large-scale deployment and cross-platform applications.

Hardware requirements

The core requirements for cloud image recognition are limited to the following two points, allowing it to run on almost any internet-connected device:

  1. Camera: For capturing images to be recognized (front/rear cameras are both suitable).
  2. Internet connectivity: Support for Wi-Fi, 4G/5G networks to communicate with cloud servers.

Applicable device range:

  • Mobile devices: Smartphones, tablets (Android/iOS).
  • Desktop devices: Windows/macOS PCs (camera required).
  • XR devices: AR/VR headsets (e.g., Apple Vision Pro).
  • Embedded devices: IoT devices with cameras and network modules (e.g., smart reading companion robots, digital signage).

OS limitations: None. Even without integrating EasyAR SDK, the cloud recognition service can be invoked via Web API, adapting to any web-supported system (e.g., Linux, Raspberry Pi OS).

Combined use with plane image tracking

Cloud recognition can be combined with local plane image tracking to enable a hybrid mode of "recognize then continuously track." To combine with plane image tracking, devices must meet:

  • EasyAR SDK integration: Applications must integrate Sense SDK and configure a License Key.
  • Hardware and platform requirements: Must meet the minimum hardware standards for plane image tracking. Refer to the Plane image tracking device support section to ensure OS version and architecture compatibility.

Use cases: Applications requiring precise post-recognition interactions (e.g., 3D model rotation, virtual button clicks).

Summary and recommendations

Cloud image recognition imposes extremely low hardware requirements and supports nearly all internet-connected devices. Combining it with plane image tracking extends functionality into a hybrid solution but requires SDK integration. Developers should choose the appropriate solution based on target scale, update frequency, and network environment.

Specifically, the hybrid solution combining cloud recognition with plane image tracking can be implemented following the guidelines in the next chapter.