Camera-Based Indoor Positioning Using Hashing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera-based indoor localization methods are computationally expensive and energy-intensive, often requiring cloud-based services that compromise privacy and are impractical for mobile devices due to high latency and energy consumption.
Innovation Solution
The Camera-Based Positioning System Using Learning (CaPSuLe) employs inexact computing and a hashing-based image matching algorithm to perform local image matching on mobile devices, reducing computational and energy costs significantly, allowing for end-to-end computation without network communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If camera-based localization is performed using traditional image matching techniques, then positioning accuracy is improved, but computational cost and energy consumption increase significantly
Solution Approach 1:
The patent segments the image matching process into two distinct phases: an offline training phase where comprehensive image databases are built and processed, and an online query phase where only hashing and comparison operations are performed. This segmentation allows computationally intensive operations to be performed offline, reducing the energy consumption during actual positioning operations on mobile devices.
Solution Approach 2:
The patent performs preliminary actions by pre-processing images during the offline training phase, including feature extraction, hashing computation, and database construction. By completing these computationally expensive operations in advance, the system minimizes the computational burden and energy consumption during online positioning operations.
Solution Approach 3:
The patent replaces traditional mechanical/computational image matching methods with a hashing-based approach. Instead of performing exhaustive pixel-by-pixel or feature-by-feature comparisons during online operations, the system uses hash functions to map images to compact representations, enabling rapid comparison with minimal computational resources.
2Productivity
If cloud-based services are used for camera-based localization, then computational resources are improved, but privacy and security are compromised
Solution Approach 1:
The patent enables mobile devices to perform self-service localization by executing the complete positioning algorithm locally on-device. The system processes images, performs hashing operations, and determines position without requiring cloud connectivity or data transmission, thereby preserving user privacy and security while maintaining computational functionality.
Solution Approach 2:
The patent segments the localization system into offline training (which can be done once on any device) and online querying (performed independently on each mobile device). This segmentation allows each device to be self-sufficient during positioning operations, eliminating the need for cloud-based processing and associated privacy risks.
3Measurement precision
If traditional image matching algorithms are used on mobile devices, then positioning accuracy is maintained, but response time increases due to computational complexity
Solution Approach 1:
The patent substitutes traditional computationally intensive image matching algorithms with a hashing-based approach. By mapping images to hash values and comparing these compact representations instead of full images or detailed features, the system achieves rapid positioning with minimal response time while maintaining acceptable accuracy levels.
Solution Approach 2:
The patent changes the parameter representation from full images or detailed feature sets to compact hash values. This parameter transformation reduces the dimensionality and complexity of comparisons, enabling fast response times on mobile devices while preserving the essential information needed for accurate positioning.
4Measurement precision
If comprehensive image processing is performed to ensure accurate matching, then positioning accuracy is improved, but memory overhead increases
Solution Approach 1:
The patent transforms the image data parameter from full-resolution images or extensive feature vectors to compact hash values. This parameter change dramatically reduces the memory required to store and process image data while maintaining the ability to perform accurate matching through hash comparison.
Solution Approach 2:
The patent extracts only the essential matching information from full images by computing hash values during the offline training phase. By storing only these extracted hash representations rather than complete images, the system reduces memory overhead while preserving the core information needed for accurate positioning.
Data Source
AI summary
A device, system, and methods are described to perform machine-learning camera-based indoor mobile positioning. The indoor mobile positioning may utilize inexact computing, wherein a small decrease in accuracy is used to obtain significant computational efficiency. Hence, the positioning may be performed using a smaller memory overhead at a faster rate and with lower energy cost than previous implementations. The positioning may not involve any communication (or data transfer) with any other device or the cloud, providing privacy and security to the device. A hashing-based image matching algorithm may be used which is cheaper, both in energy and computation cost, over existing state-of-the-art matching techniques. This significant reduction allows end-to-end computation to be performed locally on the mobile device. The ability to run the complete algorithm on the mobile device may eliminate the need for the cloud, resulting in a privacy-preserving localization algorithm by design since network communication with other devices may not be required.