HOG Descriptor Compression for Vehicle Collision Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems for collision avoidance in vehicles face challenges in efficiently extracting and transmitting histogram of oriented gradient (HOG) descriptors due to high bandwidth requirements, which can be costly and hinder real-time collision avoidance capabilities.
Innovation Solution
A system that generates a Gaussian pyramid from image frames, extracts HOG descriptors, compresses them using binarization and transformation methods, and transmits the compressed data to reduce bandwidth usage, allowing for efficient processing and collision detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If HOG descriptors are extracted using traditional techniques (Dalal-Triggs or Felzenszwalb), then contextual information can be derived for collision avoidance, but the data generated is voluminous and requires excessive bandwidth for transmission between processors
Solution Approach 1:
The patent extracts only the essential features from HOG descriptors by selecting specific bins that contain the most discriminative information for collision detection. This selective extraction reduces data volume while preserving the contextual information needed to distinguish between different object types (vehicles, pedestrians, stationary objects).
Solution Approach 2:
The patent transforms HOG descriptor data from continuous floating-point values to discrete binarized values (0 or 1). This parameter transformation dramatically reduces the amount of data that needs to be transmitted while maintaining the essential structural information needed for collision avoidance decisions.
2Measurement precision
If HOG descriptors are transmitted with high bandwidth to maintain data quality, then processing accuracy is improved, but the system cost and complexity increase
Solution Approach 1:
The patent uses simple binarization operations that can be performed with basic computational resources. This approach replaces the need for complex, high-bandwidth transmission infrastructure with simple binary data that can be transmitted over basic communication channels, reducing system complexity and cost.
3Reliability
If all HOG descriptor bins are transmitted to ensure complete information, then detection reliability is improved, but transmission time and processing delay increase
Solution Approach 1:
The patent extracts only the most critical bins from the HOG descriptor that contain essential information for collision detection. By selecting and transmitting only these key bins, the system reduces transmission time while maintaining sufficient reliability for real-time collision avoidance decisions.
Solution Approach 2:
The patent transmits a subset of HOG descriptor bins rather than all bins. This partial action approach provides just enough information for reliable collision detection without the excessive transmission time that would result from sending complete descriptor data.
Data Source
AI summary
An image processing system includes a first processor that acquires frames of image data. For each frame of data, the first processor generates a Gaussian pyramid for the frame of data, extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, compresses the HOG descriptors, and sends the compressed HOG descriptors. A second processor is coupled to the first processor and is configured to receive the compressed HOG descriptors, aggregate the compressed HOG descriptors into windows, compare data of each window to at least one stored model, and generate output based upon the comparison.