Contrast and Entropy Perception Adaptation via Probabilistic Signal Temporal Logic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Perception systems in autonomous driving and navigation suffer from perception errors, with existing methods often focusing on controlling autonomous systems rather than addressing perception errors directly, and conventional contrast adaptation methods are ineffective due to high contrast in non-object areas.
Innovation Solution
A system that uses probabilistic signal temporal logic to evaluate perception data, convert it into axioms, and adjust contrast and entropy parameters within the camera system to optimize perception, applying sharpening or smoothing filters based on entropy changes and setting histogram ranges for contrast adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If conventional contrast adaptation methods are used on the entire image, then image contrast enhancement is achieved, but object detection accuracy deteriorates due to high contrast in non-object areas
Solution Approach 1:
The patent applies local contrast adaptation specifically to detected object regions rather than the entire image. The system identifies object locations through initial detection, then applies contrast enhancement only within those specific regions using histogram equalization or contrast limited adaptive histogram equalization (CLAHE), leaving non-object areas unchanged. This resolves the contradiction by maintaining high contrast where needed for detection while avoiding false enhancements in non-object areas.
Solution Approach 2:
The patent segments the image into object regions and non-object regions based on initial detection results. By dividing the image processing into separate regions, the system can apply contrast adaptation selectively to object areas only. This segmentation approach prevents the harmful effect of applying contrast enhancement to entire image including non-object areas that would cause false detections.
2Reliability
If perception systems use formal verification methods, then perception error evaluation is improved, but system complexity increases due to probabilistic signal temporal logic implementation
Solution Approach 1:
The patent implements a feedback loop where perception results are continuously evaluated against formally verified axioms derived from probabilistic signal temporal logic. The system monitors detection outputs, compares them against expected behaviors defined in the formal logic framework, and automatically adjusts perception parameters when violations are detected. This feedback mechanism provides reliable perception error evaluation while managing complexity through automated closed-loop control rather than manual verification processes.
Solution Approach 2:
The patent uses parameter changes as a response to formal verification results. When perception errors are detected through the formal verification framework, the system automatically adjusts perception parameters such as contrast, entropy, or detection thresholds to correct the errors. This approach maintains reliability through formal verification while managing complexity by using parameter adjustment rather than complete system redesign.
Data Source
AI summary
Described is a system for contrast and entropy-based perception adaption to optimize perception. The system is operable for receiving an input image of a scene with a camera system and detecting one or more objects (having perception data) in the input image. The perception data of the one or more objects is converted into probes, which are then converted into axioms using probabilistic signal temporal logic. The axioms are evaluated based on probe bounds. If the axioms are within the probe bounds, then results are provided; however, if the axioms are outside of the probe bounds, the system estimates optimal contrast bounds and entropy bounds as perception parameters. The contrast and entropy in the camera system are then adjusted based on the perception parameters.