Object Detection via Compressed Feature Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Object detection apparatuses with low throughput or processing capability face challenges in performing both object detection and image compression encoding operations simultaneously, leading to increased processing loads and bandwidth constraints during image transmission.
Innovation Solution
The apparatus performs compression encoding on images to generate feature quantities that can be used for object detection and later decoding, reducing the processing load by transmitting these encoded features instead of raw images, allowing for efficient object detection and bandwidth management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the object detection apparatus performs both object detection operation and compression encoding operation independently, then the image can be transmitted through the communication line, but the processing load increases beyond the throughput capability of low-performance devices
Solution Approach 1:
The patent merges the object detection operation and compression encoding operation into a single integrated process. The neural network performs both tasks simultaneously by generating compressed feature representations that are directly used for object detection, eliminating the need for separate independent operations and reducing overall processing load on resource-constrained devices.
Solution Approach 2:
The neural network is designed to serve multiple functions: it performs compression encoding to reduce image data size for transmission, and simultaneously performs object detection to identify target objects. This multi-functional approach allows a single processing unit to handle both tasks that would traditionally require separate operations, optimizing resource utilization on portable terminals with limited processing capability.
2Reliability
If the object detection apparatus transmits the original image to the information processing apparatus, then the image quality is preserved, but the bandwidth consumption increases
Solution Approach 1:
The patent extracts only the essential features and compressed representations of the image that are necessary for object detection, rather than transmitting the complete original image. The neural network processes the image to generate condensed feature data that retains the critical information needed for accurate object detection while significantly reducing the data volume that needs to be transmitted through the communication line.
3Measurement precision
If the object detection apparatus uses a neural network with multiple layers for accurate object detection, then the detection precision improves, but the processing time increases
Solution Approach 1:
The neural network performs preliminary compression and feature extraction during the encoding phase, preparing the data in advance for efficient detection. By pre-processing the image to extract relevant features and compress the data representation before transmission, the system reduces the computational burden during the actual detection phase, thereby decreasing processing time while maintaining detection precision.
Data Source
AI summary
An object detection apparatus includes: a generation unit that performs compression encoding on each of a first image obtained from an image generation apparatus and a second image indicating a detection target object so as to extract a feature quantity that allows object detection and so as to be decoded later, thereby generate respective one of first encoding information that is the compressed, encoded first image and that is usable as a first feature quantity that is the feature quantity of the first image, and second encoding information that is the compressed, encoded second image and that is usable as a second feature quantity that is the feature quantity of the second image; and a detection unit that detects the detection target object in the first image, by using the first and second feature quantities.