Selective Object Detection Using Pixel Delta Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection algorithms in video frames are resource-intensive and time-consuming, particularly in mobile and IoT devices, and often require significant CPU and memory resources, making them unsuitable for real-time video streaming and efficient resource management.
Innovation Solution
A method that selectively applies the object detection algorithm to video frames by analyzing pixel differences using codec metadata, allowing for the prediction of bounding boxes in frames with low pixel delta, thereby reducing the number of times the algorithm needs to be applied and conserving resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If object detection algorithm is applied to every frame, then detection precision is maintained, but resource consumption increases significantly
Solution Approach 1:
The patent implements frame skipping by comparing pixel delta between consecutive frames and skipping the object detection algorithm for frames with low pixel differences (below threshold). This allows the system to maintain detection precision for significant changes while skipping redundant processing for similar frames, directly resolving the contradiction between precision and resource consumption.
Solution Approach 2:
The patent introduces a pixel delta threshold parameter that dynamically controls whether to apply the object detection algorithm. By changing the processing parameter (apply algorithm or skip) based on the pixel delta value, the system optimizes resource usage while maintaining detection precision when necessary.
2Reliability
If object detection algorithm is applied to every frame, then object tracking continuity is ensured, but processing time increases
Solution Approach 1:
The patent skips object detection processing for frames with pixel delta below the threshold, reducing processing time significantly. The bounding box from the previous frame is carried forward for skipped frames, maintaining tracking continuity without redundant processing, thus resolving the contradiction between reliability and time loss.
Solution Approach 2:
The patent performs preliminary pixel delta comparison before applying the full object detection algorithm. This preliminary action determines whether the expensive detection algorithm is necessary, allowing the system to prepare and carry forward bounding box information in advance for frames that don't require full processing.
3Measurement precision
If object detection algorithm is applied to every frame, then detection accuracy is maintained, but computational load increases
Solution Approach 1:
The patent skips the computationally intensive object detection algorithm for frames with low pixel delta values, directly increasing frame processing rate. The system maintains detection accuracy for frames that truly contain changes while rapidly processing or skipping frames with minimal changes, resolving the contradiction between accuracy and productivity.
Solution Approach 2:
The patent applies partial processing by performing only the lightweight pixel delta comparison for all frames, and only applying the full object detection algorithm when necessary (when pixel delta exceeds threshold). This partial action approach maintains accuracy where needed while maximizing processing rate overall.
Data Source
AI summary
An object detection algorithm is selectively applied to frames in a video. A frame in the video is analyzed using a set of neural networks of the object detection algorithm to detect a location of an object in the frame and predict a bounding box for the object in the frame. A magnitude of a delta between the frame and a second frame is determined. The magnitude of the delta is determined based on a difference between the two frames in values of at least one parameter of their respective sets of pixels. Responsive to the magnitude of the delta being less than a threshold, a new bounding box is predicted for the object in the second frame without analyzing the second frame using the set of neural networks.