Machine-Learned Photographic ROI Beyond Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image capture devices often rely on automated functions like automatic exposure and focus, which may not optimize image quality due to the use of object regions of interest detected by object detectors, as these regions may not be ideal for photographic purposes.
Innovation Solution
A computing device employs a trained machine learning algorithm to determine a photographic region of interest based on the input image and object region of interest, optimizing parameters such as skin coverage and edge maximization for improved image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If object detectors are used to determine regions of interest for automated photographic functions, then the system can operate automatically without manual input, but the determined regions may not be optimal for photographic quality
Solution Approach 1:
A machine learning model is introduced as an intermediary between the object detector and the photographic functions. The object detector first identifies object regions, which are then processed by the machine learning model to generate optimized photographic regions of interest. This intermediary layer transforms the raw object detection results into photographically optimal regions, resolving the contradiction between automation and quality.
2Device complexity
If traditional object detectors are used to identify regions of interest, then the system complexity remains low, but the photographic function performance is suboptimal
Solution Approach 1:
The system changes the parameters used for region determination by training a machine learning model on specific photographic parameters such as skin coverage, edge maximization, and photographic function optimization. This parameter transformation allows the system to maintain relatively simple architecture while significantly improving photographic function performance through optimized parameter selection.
Data Source
AI summary
Apparatus and methods related to photography are provided. A computing device can receive an input image. An object detector of the computing device can determine an object region of interest of the input image that is associated with an object detected in the input image. A trained machine learning algorithm can determine an output photographic region of interest for the input image based on the object region of interest and the input image. The machine learning algorithm can be trained to identify an output photographic region of interest that is suitable for use by a photographic function for image generation. The computing device can generate an output related to the output photographic region of interest.