Image Stream Object Masking Using Discarded Pixel Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection algorithms fail to properly identify and locate objects that are partially visible or suddenly appear in an image stream, leading to unreliable masking or marking, particularly near image stream boundaries, which compromises privacy protection.
Innovation Solution
A method and device that generate output image streams by processing input streams while retaining and utilizing discarded pixel information about the scene, including lower quality and non-rectangular pixels, to enhance object detection and tracking, even for partially visible objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If object detection algorithms use only the pixel information included in the output image stream, then the processing is simple and fast, but objects that are partially visible or suddenly appear near boundaries cannot be properly detected
Solution Approach 1:
The patent applies preliminary action by performing object detection on discarded pixel information before the main output image stream processing. This allows the system to detect objects that are partially visible or about to enter the scene, so that masking can be applied proactively when these objects become visible in the output stream, rather than reacting after detection failure.
Solution Approach 2:
The patent segments the pixel information processing into two separate pathways: one pathway processes pixels that will be included in the output image stream, while another pathway processes discarded pixel information separately for detection purposes. This segmentation allows the system to use different processing strategies for different pixel sets, improving detection reliability without overwhelming the main processing pipeline.
2Reliability
If the object detection algorithm requires multiple updates before tracking can begin, then tracking stability is improved, but objects entering or leaving the scene cannot be tracked in time for proper masking
Solution Approach 1:
The system performs preliminary object detection and tracking initialization using discarded pixel information before the object fully enters the output image stream. This allows the tracking algorithm to be pre-initialized with object parameters (position, velocity, size) so that when the object becomes visible in the output stream, masking can begin immediately without waiting for multiple detection updates.
Solution Approach 2:
The discarded pixel information acts as an intermediary that bridges the gap between object appearance and reliable detection. By using this intermediate data source, the system can initialize tracking earlier and provide advance notice to the masking system, effectively reducing the time delay between object entry and masking activation.
3Reliability
If discarded pixel information is used for object detection, then detection of partially visible objects is improved, but additional processing resources are required
Solution Approach 1:
The patent applies local quality by focusing computational resources on processing discarded pixel information specifically for detection purposes, rather than processing all pixels with the same level of detail. The system can apply simpler, faster detection algorithms to the discarded pixel set since this is only one component of the overall processing pipeline, and the results are used specifically for early object detection and masking initialization.
Data Source
AI summary
A method of masking or marking an object in an image stream is provided, including: generating one or more output image streams by processing an input image stream capturing a scene, including discarding pixel information about the scene provided by pixels of the input image stream, such that the discarded pixel information about the scene is not included in any output image stream; and detecting an object in the scene using the discarded pixel information, wherein generating the one or more output image streams includes masking or marking the detected object in at least one output image stream once deciding that the object is at least partially visible within this at least one output image stream. A corresponding device, computer program and computer program product are also provided.