Unattended Object Detection via Occlusion-Based ML Policy Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated detection of unattended objects in crowded environments is challenging due to occlusion and the need for efficient computational methods, as deep learning techniques are computationally expensive while simple feature analysis may not be sufficient in crowded or occluded conditions.
Innovation Solution
A feature matrix policy selector determines the appropriate machine learning policy based on occlusion levels, selecting between deep learning and simple feature analysis techniques to identify unattended objects, balancing computational resources with detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep learning techniques are used for object detection, then detection accuracy is improved, but computational cost increases
Solution Approach 1:
The system dynamically selects between deep learning and simple feature analysis techniques based on real-time occlusion level assessments. When occlusion is high, deep learning models are deployed for accurate detection despite higher computational cost. When occlusion is low, simpler and faster feature analysis methods are used, reducing computational overhead while maintaining sufficient accuracy.
Solution Approach 2:
The system changes the parameter of machine learning technique selection based on the occlusion parameter. By evaluating occlusion levels and adjusting the detection method accordingly, the system optimizes the balance between detection accuracy and computational resource consumption for different environmental conditions.
2Productivity
If simple feature analysis is used for object detection, then computational efficiency is improved, but detection accuracy deteriorates in crowded or occluded conditions
Solution Approach 1:
The system dynamically adjusts the detection technique based on occlusion assessment. In crowded or occluded environments where simple feature analysis would fail, the system transitions to deep learning techniques that can accurately detect unattended objects despite the increased computational requirements.
3Device complexity
If a single machine learning model is used for all conditions, then system complexity is reduced, but adaptability to different occlusion levels deteriorates
Solution Approach 1:
The system implements a universal policy selection mechanism that can adaptively choose between multiple machine learning models (deep learning and simple feature analysis) based on occlusion conditions. This multi-functional approach allows a single system to handle varying environmental conditions effectively without requiring separate dedicated systems for each scenario.
Data Source
AI summary
Techniques for unattended object detection using machine learning are disclosed. A machine learning policy, for use in identifying unattended objects in a captured image depicting one or more objects, is generated. The generating includes determining a level of occlusion in the captured image relating to the objects, and determining the machine learning policy based on the determined level of occlusion. A machine learning model is selected, from among a plurality of pre-defined machine learning models, based on the generated machine learning policy. An unattended object is detected in the captured image using the selected machine learning model.