Object Detection via Aggregated Probability Scores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection systems based on machine learning models have limited reliability and accuracy, especially for small or distant objects, as they rely solely on image analysis without considering physical attributes and movement patterns, leading to reduced classification accuracy and robustness.
Innovation Solution
A computer-implemented method and system that aggregates probability scores from machine learning models with physical attributes and movement patterns analysis, using neural networks to enhance object detection by computing and combining first, second, and third probability scores based on image analysis, physical attributes, and movement patterns, respectively, to improve classification accuracy and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning models are used for object detection based solely on image analysis, then object detection can be performed, but classification accuracy and reliability are limited especially for small or distant objects
Solution Approach 1:
The patent combines three different detection approaches: machine learning-based object detection, physical attributes detection, and movement patterns detection. By merging these multiple detection methods and aggregating their probability scores, the system achieves higher classification accuracy and reliability than any single method alone, particularly for small or distant objects
Solution Approach 2:
The patent introduces probability score aggregation as an intermediary mechanism that combines results from multiple detection sources. The aggregation module acts as a mediator that synthesizes ML model outputs, physical attributes analysis, and movement patterns analysis to produce a final aggregated probability score that reflects overall detection confidence
2Reliability
If complex machine learning models are used to improve detection accuracy, then classification reliability increases, but computational resources and system complexity increase
Solution Approach 1:
The patent segments the object detection task into three independent modules: ML-based detection, physical attributes detection, and movement patterns detection. Each module operates independently and contributes its own probability score, allowing the system to achieve high reliability without requiring a single complex model. This segmentation reduces overall system complexity while maintaining or improving accuracy
3Measurement precision
If high-performance imaging sensors and complex models are used to achieve accurate detection, then detection quality improves, but computational resources and cost increase
Solution Approach 1:
The patent applies partial action by using multiple detection methods only to the extent necessary for each object type and scenario. The system aggregates probability scores from ML models, physical attributes, and movement patterns selectively, achieving sufficient detection accuracy without applying all computational resources to every detection case. This allows accurate detection with lower computational overhead compared to using a single high-performance model on all inputs
Data Source
AI summary
Presented herein are systems and methods for increasing reliability of object detection, comprising, receiving a plurality of images of one or more objects captured by imaging sensor(s), receiving an object classification coupled with a first probability score from machine learning model(s) trained to detect the object(s) and applied to the image(s), computing a second probability score for classification of the object(s) according to physical attribute(s) of the object(s) estimated by analyzing the image(s), computing a third probability score for classification of the object(s) according to a movement pattern of the object(s) estimated by analyzing at least some consecutive images, computing an aggregated probability score aggregating the first, second and third probability scores, and outputting, in case the aggregated probability score exceeds a certain threshold, the classification of each object coupled with the aggregated probability score for use by object detection based system(s).