Mobile Camera Focus via Object Recognition Cues
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile devices with passive auto-focusing struggle to quickly and accurately focus on objects of interest, especially in blurred images, often requiring users to manually reorient the device or tap the screen, leading to delays and potential misfocus.
Innovation Solution
Implementing an objects of interest recognition system that processes image data to identify candidate objects based on cues like color contrast, edge density, and superpixel straddling, allowing passive auto-focusing to be performed at the selected object location, with a graphical symbol indicating the focus point, which can be moved dynamically as the device is reoriented or new objects become relevant.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If passive auto-focusing is used to determine focus without user intervention, then ease of operation is improved, but measurement precision deteriorates because it cannot accurately identify objects of interest in blurred images
Solution Approach 1:
The system performs preliminary object detection and identification before the focusing operation. By analyzing image content to identify objects of interest and their locations in advance, the system establishes a target focus area before the passive auto-focus process begins, ensuring that the subsequent focusing operation is directed at the correct object rather than relying solely on intensity differences in blurred images.
Solution Approach 2:
The system introduces an intermediary object detection and recognition module between the image capture and passive auto-focus processes. This intermediary analyzes the captured image to identify objects of interest, determines their locations, and provides focus area guidance to the passive auto-focus system, bridging the gap between automatic operation and precise focusing.
2Device complexity
If the focus area is fixed at the center of the display, then device complexity is reduced, but adaptability deteriorates because the object of interest may not be in the focus area
Solution Approach 1:
The system makes the focus area dynamic by continuously or periodically adjusting it based on detected objects of interest. Instead of a fixed center focus area, the system dynamically determines the focus area location by identifying objects in the image, allowing the focus area to adapt to different scenes and objects automatically.
Solution Approach 2:
The system changes the parameters of the focus area (position, size, location) based on object detection results. By analyzing image content and identifying objects of interest, the system adjusts the focus area parameters to match the detected objects, enabling the focus area to adapt to different scenarios without increasing overall device complexity.
3Measurement precision
If passive auto-focusing captures and evaluates images at different lens positions, then measurement precision is improved, but loss of time increases due to the evaluation process
Solution Approach 1:
The system performs preliminary object detection and focus area determination before the passive auto-focus evaluation process. By identifying objects of interest and establishing the target focus area in advance, the system guides the subsequent lens position evaluation process, reducing the number of positions that need to be evaluated and accelerating the overall focusing process.
Solution Approach 2:
The system implements feedback by using object detection results to guide the passive auto-focus process. The detected object locations and characteristics provide feedback that informs which lens positions should be evaluated and how to interpret the sharpness measurements, enabling the system to converge on the correct focus more quickly.
Data Source
AI summary
An approach is provided for acquiring images with camera-enabled mobile devices using objects of interest recognition. A mobile device is configured to acquire an image represented by image data and process the image data to identify a plurality of candidate objects of interest in the image. The plurality of candidate objects of interest may be identified based upon a plurality of low level features or “cues” in the image data. Example cues include, without limitation, color contrast, edge density and superpixel straddling. A particular candidate object of interest is selected from the plurality of candidate objects of interest and a graphical symbol is displayed on a screen of the mobile device to identify the particular candidate object of interest. The particular candidate object of interest may be located anywhere on the image. Passive auto focusing is performed at the location of the particular candidate object of interest.