Electronic Camera Object Tracking with Segmented Feature Update
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional object tracking methods in electronic cameras fail to maintain accurate tracking when the object changes appearance or is not visible, leading to mistracking and loss of the target due to inclusion of background features in the update area.
Innovation Solution
The electronic camera determines an update area using specific features for stable object detection and corrects displacement, allowing continuous tracking by updating the feature amount when significant changes occur, and automatically terminates tracking when it becomes impossible to continue, displaying appropriate feedback to the user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional object tracking methods update the feature amount using the entire tracking area, then the tracking can be maintained, but background features are included in the update area causing mistracking when the object changes appearance
Solution Approach 1:
The patent segments the tracking area into a valid update area and an invalid update area based on feature amount evaluation. The valid update area contains only the object of interest, while the invalid update area contains background regions. This segmentation prevents background features from contaminating the feature amount update, resolving the contradiction between maintaining tracking and avoiding mistracking.
Solution Approach 2:
The patent applies different quality criteria to different regions within the tracking area. By evaluating feature amounts locally across the tracking area and identifying regions with significant feature differences (valid update area) versus regions with minimal feature differences (invalid update area), the system ensures that only high-quality, object-specific features are used for updating, thereby maintaining both tracking reliability and feature precision.
2Adaptability or versatility
If the tracking area is expanded to include more of the object, then more features are available for tracking, but background features are also included causing loss of target when object is not visible
Solution Approach 1:
The patent segments the expanded tracking area into valid and invalid update areas based on local feature evaluation. Even when the tracking area is expanded to maintain robustness during object occlusion or appearance changes, the segmentation ensures that only regions containing actual object features (valid update area) contribute to the feature amount update, while background regions (invalid update area) are excluded. This resolves the contradiction by allowing expanded tracking coverage without compromising target acquisition accuracy.
3Extent of automation
If face detection is used for alignment in auto focus and auto exposure, then automatic focusing on objects is supported, but it cannot detect faces when not visible or track other objects such as pets
Solution Approach 1:
The patent transforms the specialized face detection system into a universal object tracking system. By using generic feature amount evaluation and tracking algorithms that work with any distinguishable object features rather than face-specific features, the system maintains automatic focusing support while gaining the ability to track any object including pets and other non-face targets, thereby achieving both automation and versatility.
Solution Approach 2:
The patent changes the detection parameters from face-specific features to generic object features. By evaluating feature amounts based on general image characteristics rather than face-specific patterns, the system can detect and track any object type. This parameter change enables the system to maintain automatic focusing functionality while expanding object detection range to include pets and other objects that were previously undetectable.
Data Source
Figure 1
Figure 2
Figure 3(a)~3(e)
AI summary
To provide an electronic camera and an image processing method which enable highly accurate tracking processing even in the case of an object changing appearance. The electronic camera includes: a tracking processing unit (137) which performs a search in a frame using a color feature amount of an object to be tracked, calculates a first evaluation value indicating a degree of matching between an image of the search result and the object, and determines a target area estimated to include the image of the object based on the first evaluation value; an update detection circuit (139) which performs a search in the frame using a shape feature amount of the object to be tracked, calculates a second evaluation value indicating a degree of matching between an image of the search result and the image of the object, and determines an update area for updating the first feature amount based on the second evaluation value; an update determination circuit (140) which determines whether or not to update the first feature amount, depending on whether or not the first evaluation value or the second evaluation value satisfies a predetermined condition; and a tracking result drawing circuit (144) draws an outer frame of the target area when determined not to update, and draws an outer frame of the update area when determined to update.