Differences between EasyAR dense spatial mapping and ARKit/ARCore

This chapter introduces the differences between EasyAR dense spatial mapping and similar features in ARKit/ARCore (generating dense mesh or depth map), and compares the advantages and disadvantages of using mesh versus depth map for achieving realistic AR effects.

EasyAR versus ARKit-generated mesh

EasyAR dense mapping reconstructs the environment's mesh in real-time for occlusion, collision and other effects. ARKit provides similar scene reconstruction functionality, allowing export of environmental meshes. In principle, ARKit utilizes both RGB cameras and LiDAR depth cameras, thus it only supports devices from iPad Pro 4th generation onward. EasyAR dense mapping relies solely on RGB cameras, offering broader device compatibility.

Achieving realistic AR effects: using mesh or depth map

Google ARCore does not support generating environmental mesh information. Instead, it provides a per-frame depth map to achieve occlusion effects.

Apple ARKit offers collision and occlusion effects by accessing depth data from rear cameras and the scene through similar depth map (SceneDepth) methods.

Both depth maps and meshes can enhance reality for realistic occlusion and collision effects. Each has distinct strengths and weaknesses, compared as follows:

	Depth Map	Mesh
Representation Dimension	2.5D	3D
Update Method	Real-time per frame	Incremental/batch reconstruction
Dynamic Object Support	Good	Poor
Cross-view Consistency	Poor	Good
Persistence Capability	Not supported	Supported