Distributed Object Detection via Ad-hoc Node Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object detection and tracking technologies face challenges in resource-constrained environments, such as mobile devices with limited computational power and battery life, making it difficult to perform real-time object detection and tracking without access to powerful GPUs or TPUs.
Innovation Solution
The implementation of distributed computation across multiple nodes in an ad-hoc network, where object detection and tracking tasks are distributed among devices based on their available computational resources and battery life, allowing for opportunistic and efficient use of resources without the need for centralized data centers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If object detection is performed using CNN-based algorithms on mobile devices, then detection accuracy is improved, but power consumption and computational load increase significantly
Solution Approach 1:
The patent segments the object detection task into multiple frames and distributes them across different nodes in an ad-hoc network. Instead of processing all frames on a single mobile device, the system divides the computational workload by sending different video frames to different nodes for detection, thereby reducing the power consumption and computational load on individual devices while maintaining detection accuracy through aggregated results.
2Ease of operation
If object detection is performed on battery-powered devices with standard CPUs, then device portability and accessibility are improved, but detection speed and real-time performance deteriorate
Solution Approach 1:
The patent merges the computational capabilities of multiple mobile devices in an ad-hoc network to achieve real-time object detection. By combining the processing power of several devices with standard CPUs, the system achieves detection speeds comparable to powerful GPUs while maintaining device portability and accessibility. The distributed architecture allows parallel processing across multiple nodes, significantly improving overall detection speed.
3Use of energy by moving object
If distributed computation is used across multiple nodes, then power consumption per device is reduced, but system complexity and coordination overhead increase
Solution Approach 1:
The patent implements a self-service mechanism where each node in the ad-hoc network autonomously performs object detection on assigned frames and contributes results to the overall detection process. Nodes independently manage their own computational resources and make local decisions about frame processing, reducing the need for centralized coordination. This self-organizing approach minimizes system coordination complexity while effectively distributing power consumption across multiple devices.
Data Source
AI summary
Techniques for distributed object detection and tracking are described. In an example method, a first current frame from a series of frames is sent to a first node, for detection of a first object. After object detection information for the first object is received in return, a second node is selected, and a second current frame is sent to the second node for an updated object detection. In addition, while waiting for the results of the updated object detection, two or more frames following the second current frame are sent to respective tracking nodes. Object modelling information indicating location and/or classification of one or more objects, as derived from the previously received object detection information is also sent to each of the respective tracking nodes. Tracking information for the first object is received from each of the respective tracking nodes.