Subject Tracking Unit for Automatic Image Capture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies, such as those described in Patent Literature 1, fail to automatically record moments other than facial expressions, like a person falling or a child stopping movement, and cannot apply to subjects without facial expressions.
Innovation Solution
An image processing apparatus that includes a subject tracking unit capable of detecting and tracking any subject based on saliency maps, allowing for automatic shutter activation when significant changes in the subject's state or position occur, regardless of facial expressions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If facial expression detection is used to trigger automatic recording, then the system can automatically capture smiling moments, but it cannot detect other important moments such as when a person falls or when movement stops
Solution Approach 1:
The subject tracking unit is designed to track any moving object in the image regardless of its type or state. Instead of relying on facial expression detection which only works for specific scenarios, the system uses general object tracking that can detect changes in position, size, and shape of any subject including falling persons, running animals, or stationary objects, making the automatic recording function universally applicable to diverse situations
Solution Approach 2:
The system monitors changes in tracking parameters such as subject position coordinates, subject size (area), and subject shape characteristics. When these parameters change beyond predetermined thresholds, it triggers recording. This parameter-based approach replaces expression-based detection, enabling the system to capture various moments including falls (sudden position change), stops (velocity change), and other significant events
2Measurement precision
If facial expression detection is used, then the system can identify specific emotional moments, but it cannot be applied to subjects without facial expressions such as animals or inanimate objects
Solution Approach 1:
The subject tracking unit employs general object detection and tracking algorithms that work on any visible object in the scene without requiring facial features. The system identifies subjects based on their visual characteristics and tracks their motion, enabling it to monitor animals, vehicles, sports equipment, or any other movable object, thus achieving universal applicability across different subject types
Solution Approach 2:
The invention extracts the essential tracking functionality from the facial expression detection framework. Instead of relying on face-specific features, it extracts and tracks fundamental object properties such as position, velocity, acceleration, and shape changes, which are applicable to all moving objects regardless of whether they have faces or not
3Device complexity
If the system only triggers on facial expressions, then it can maintain simple processing logic, but it misses important moments like when a running person falls or a child stops moving
Solution Approach 1:
The control unit monitors multiple tracking parameters including subject position (x, y coordinates), subject size (area), and derived velocity and acceleration. When any of these parameters change beyond predetermined thresholds, the system triggers recording. This parameter-based approach maintains relatively simple processing logic while significantly improving the reliability of capturing important moments such as falls, stops, and other dynamic events
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image processing apparatus, method and non-transitory computer program storage device cooperate to process successive images. Respective frames are created and positioned within the successive images, where each frame has a border. When changes between the frame borders are detected, a controller triggers the capturing of an image. This approach results in the capturing of interesting moments, even if the subject is not a human subject. The change in frame boundaries may be categorized in a variety of ways, including change in aspect ratio, shape, orientation, and position, for example. By detecting the changes in this way, an imaging device can capture images of interesting events automatically.