Forward-Backward Motion History Accumulation for Moving Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion detection methods for moving objects in video sequences, especially in surveillance and reconnaissance, face challenges due to errors in image registration and unclear object boundaries, leading to over- or partial-detections which can fail subsequent object classification tasks.
Innovation Solution
A system utilizing forward and backward motion history accumulation, with multiple detection methods (Methods 1-4) that calculate frame differences and apply thresholds to accurately detect moving objects, incorporating adaptive parameter optimization based on input image characteristics for improved detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of frames used for motion accumulation is increased to detect slow moving objects, then detection sensitivity is improved, but motion areas are increased by adding the area of the object's previous or next locations
Solution Approach 1:
The patent segments the motion detection process into forward accumulation (comparing current frame with previous frames) and backward accumulation (comparing current frame with future frames). By processing motion history in two separate directional passes and then combining results, the method isolates the object's true location from ghost detections caused by motion blur across multiple frames.
Solution Approach 2:
The patent applies backward motion accumulation by comparing the current frame with future frames, which is the reverse of conventional forward-only accumulation. This inversion allows the system to distinguish between objects that are truly present in the current frame versus objects that have moved into or out of the frame, thereby reducing ghost detections while maintaining sensitivity to slow-moving objects.
2Ease of operation
If a fixed threshold value is used for motion detection, then processing simplicity is maintained, but object boundaries are not tight and clear
Solution Approach 1:
The patent implements adaptive thresholding where the threshold value is dynamically adjusted based on local image characteristics and motion history accumulation results. Instead of using a fixed global threshold, the system adapts thresholds to local variations in brightness, texture, and motion patterns, enabling tight and clear object boundaries while maintaining processing efficiency through localized adaptations rather than complex global optimization.
3Measurement precision
If image registration is performed to stabilize video, then motion detection accuracy is improved, but registration errors still affect detection precision
Solution Approach 1:
The patent performs image registration as a preliminary step before motion detection, stabilizing the video sequence to remove camera motion. By pre-aligning all frames to a reference frame through registration, the system establishes a consistent coordinate system that improves subsequent motion detection accuracy, while the forward-backward accumulation method then compensates for any remaining registration errors.
Solution Approach 2:
The patent uses bidirectional motion history accumulation that inherently provides feedback mechanisms to identify and correct registration errors. By comparing each frame with both previous and future frames, the system can detect inconsistencies caused by registration errors and adjust accordingly, reducing their impact on final detection precision through the consensus of multiple temporal comparisons.
Data Source
AI summary
Described is a system for detecting moving objects using multi-frame motion history images. An input video sequence of consecutive registered image frames is received. The sequence of consecutive registered image frames comprises forward and backward registered image frames registered to a coordinate system of a reference image frame. Frame differences are computed between each of the consecutive registered image frames and the reference image frame. The frame differences are accumulated based on characteristics of the input video sequence to compute a corresponding motion response value. A selected threshold value is then applied to the motion response value to produce at least one binary image used for detection of moving objects in the input video sequence. Additionally, the invention includes a system for adaptive parameter optimization by input image characterization, wherein parameters that are based on characteristics of the image influence the motion detection process.