Apple’s depth-based imaging relies on a combination of hardware sensors and computational models to estimate per-pixel distance from the camera. This is achieved through techniques such as stereo disparity (dual cameras), structured light (e.g., TrueDepth system), and LiDAR (time-of-flight measurement using infrared pulses). Depth maps are generated by calculating the time delay or disparity between emitted and received signals, producing a dense representation of scene geometry. Machine learning models then refine this data using semantic segmentation to distinguish subjects (faces, objects) from the background. For dynamic scene effects (such as parallax or motion-based perspective shifts), the system uses the depth map to reproject the image into a pseudo-3D space. When the phone moves, the rendering engine adjusts the viewpoint slightly, simulating real-world perspective changes. This involves real-time transformations using techniques from computer vision and graphics, includin...