Trackable Privacy Masks for Video Streams Without Information Loss
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video stream masking technologies completely obscure identifiable features, leading to loss of information necessary for tracking and analysis in applications like surveillance and factory automation, thus posing a challenge in balancing privacy and functionality.
Innovation Solution
Generate trackable privacy masks that encode unique feature vectors for objects, allowing them to be tracked across images while maintaining privacy by transforming visual features into a lower dimensional domain, which can be decoded for analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional masking is applied to obscure objects in video streams, then privacy protection is improved, but tracking capability and information availability deteriorate
Solution Approach 1:
The patent segments the masking process into two distinct components: a privacy mask that obscures identifiable features and a feature vector that preserves tracking information. The mask and feature vector are generated separately and then combined, allowing privacy protection and tracking functionality to coexist without interfering with each other.
Solution Approach 2:
The patent introduces feature vectors as an intermediary representation that bridges the gap between privacy protection and tracking capability. Instead of directly using obscured mask pixels for tracking, the system extracts meaningful features from the original image, encodes them into compact vectors, and uses these vectors for tracking while the mask handles privacy protection.
2Object-affected harmful factors
If visual features are completely obscured for privacy, then privacy protection is improved, but object recognition and analysis capability deteriorate
Solution Approach 1:
The patent extracts essential tracking and recognition features from the original image before applying the privacy mask. By taking out the critical information (feature vectors) prior to obscuration, the system ensures that privacy protection does not compromise the ability to recognize and analyze objects.
Solution Approach 2:
The patent transforms visual features from pixel space into a different parameter space (feature vectors) that preserves discriminative information while being compact and suitable for tracking. This parameter transformation allows the system to maintain recognition accuracy without relying on visible visual features.
3Object-affected harmful factors
If masks are added to video streams, then privacy protection is improved, but processing complexity and computational load increase
Solution Approach 1:
The patent performs feature extraction and encoding as preliminary actions before the masking process. By preparing feature vectors in advance, the system avoids the need for complex real-time processing of masked images for tracking purposes, thereby reducing overall computational complexity.
Solution Approach 2:
The patent creates a compact copy of essential features in the form of feature vectors. Instead of processing the full-resolution masked images for tracking and analysis, the system uses these compressed feature vector copies, which significantly reduces computational requirements while maintaining tracking accuracy.
Data Source
AI summary
An approach for providing a privacy mask for video streams. The approach involves, for example, processing one or more images of a video stream to detect at least one object. The approach also determines one or more feature vectors from a plurality of pixels of the one or more images that correspond to the at least one detected object. The approach further involves encoding the one or more feature vectors into at least one mask that obscures the at least one object in the one or more images. The at least one mask is trackable across the one or more images of the video stream based on the one or more feature vectors. The approach further involves causing an addition of the at least one mask to the one or more images of the video stream and providing the video stream with the at least one mask as an output.