Neural Network Object Localization Using Fiducial Lines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensing systems face challenges in efficiently identifying and locating objects in image data, particularly in real-time applications, due to high computational requirements and resource constraints.
Innovation Solution
The method involves adding fiducial lines to images and using a deep neural network (DNN) optimized with a genetic algorithm to determine the relative location of objects, thereby reducing computational resources and time required for object detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a deep neural network is used to determine object locations in image data, then object detection accuracy is improved, but computational resource requirements and processing time increase
Solution Approach 1:
The patent applies segmentation by dividing the image processing task into distinct stages: first processing fiducial markers to establish a coordinate system, then using this transformed coordinate system to locate objects. This segmentation allows the complex object detection problem to be broken down into simpler, more efficient sub-tasks that reduce overall computational burden while maintaining accuracy
Solution Approach 2:
The patent implements preliminary action by first processing and transforming fiducial markers before using them for object location. The fiducial markers are detected and used to create a transformed coordinate system in advance, which then facilitates faster and more efficient object localization without requiring the full neural network to process the entire image from scratch
2Measurement precision
If a deep neural network is used to determine object locations in image data, then object detection accuracy is improved, but computational resource requirements increase
Solution Approach 1:
The patent applies segmentation by dividing the image processing task into distinct stages: first processing fiducial markers to establish a coordinate system, then using this transformed coordinate system to locate objects. This segmentation allows the complex object detection problem to be broken down into simpler, more efficient sub-tasks that reduce overall computational burden while maintaining accuracy
Solution Approach 2:
The patent extracts and utilizes fiducial markers as separate reference elements from the main image data. By detecting these markers first and using them to transform the coordinate system, the system separates the reference frame establishment from object detection, thereby reducing the computational resources needed for the overall task while maintaining detection accuracy
3Speed
If real-time object detection is implemented, then system responsiveness is improved, but computational complexity increases
Solution Approach 1:
The patent implements preliminary action by first processing and transforming fiducial markers before using them for object location. The fiducial markers are detected and used to create a transformed coordinate system in advance, which then facilitates faster and more efficient object localization without requiring the full neural network to process the entire image from scratch
Solution Approach 2:
The patent applies segmentation by dividing the image processing task into distinct stages: first processing fiducial markers to establish a coordinate system, then using this transformed coordinate system to locate objects. This segmentation allows the complex object detection problem to be broken down into simpler, more efficient sub-tasks that reduce overall computational burden while maintaining accuracy
Data Source
AI summary
A location of a first object can be determined in an image. A line can be drawn on the first image based on the location of the first object. A deep neural network can be trained to determine a relative location between the first object in the image and a second object in the image based on the line. The deep neural network can be optimized by determining a fitness score that divides a number of deep neural network parameters by a performance score. The deep neural network can be output.