Smartphone Portrait Mode via AR Point Cloud Depth Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional smartphones struggle to achieve effective portrait mode effects due to the absence of adjustable apertures and the need for additional depth cameras, which increase costs and design challenges, while existing methods for simulating background blur are imperfect and prone to subject edge blurring.
Innovation Solution
Utilizing a single camera to access and leverage depth information from an augmented reality Point Cloud to perform depth-based imaging effects, such as portrait mode, without the need for a second camera, by determining the device's location and accessing precise 3D depth data stored in the Point Cloud to apply blur and enhance image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a second depth camera is added to achieve portrait mode effects, then imaging functionality is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses AR Point Cloud data to create a virtual copy of the real-world depth information. Instead of physically capturing depth with a second camera, the system copies depth data from pre-stored Point Cloud representations of environments, allowing portrait mode effects without additional hardware cameras.
Solution Approach 2:
The patent introduces AR Point Cloud data as an intermediary between the single camera and the depth information needed for portrait mode. The Point Cloud acts as a mediator that provides depth data without requiring a physical depth-sensing camera, resolving the contradiction between functionality and hardware complexity.
2Adaptability or versatility
If a second depth camera is added to achieve portrait mode effects, then imaging functionality is improved, but manufacturing cost increases
Solution Approach 1:
The system copies depth information from AR Point Cloud databases rather than physically manufacturing a second depth camera. This virtual copying approach eliminates the need for additional camera hardware, reducing manufacturing costs while maintaining portrait mode functionality.
Solution Approach 2:
The patent makes the single camera system multi-functional by combining it with AR Point Cloud data. The same camera serves both standard imaging and depth-based portrait mode functions, eliminating the need for dedicated depth camera hardware and reducing manufacturing complexity and cost.
3Adaptability or versatility
If conventional methods are used to simulate background blur, then portrait mode effect is achieved, but image quality deteriorates due to subject edge blurring
Solution Approach 1:
The patent copies precise depth boundaries from the AR Point Cloud data to accurately define subject edges. This virtual depth map provides clean, precise boundaries that prevent edge blurring, unlike conventional simulation methods that rely on less accurate depth estimation algorithms.
Solution Approach 2:
The patent replaces the mechanical/optical depth-sensing system (second camera) with a data-processing approach using AR Point Clouds. This substitution provides more precise depth information for edge detection, improving image quality by preventing subject edge blurring while achieving portrait mode effects.
Data Source
AI summary
One embodiment provides a method, including: determining, using a processor, a location of an information handling device; providing, using a camera of the information handling device, a camera view encompassing at least a portion of the location, wherein the portion encompassed by the camera view comprises at least one object; accessing a data store comprising depth information for the at least one object in the location; and leveraging the depth information to produce an imaging effect on the at least one object in the camera view of the portion. Other aspects are described and claimed.