Hybrid Image Processing Assembly for Bandwidth-Constrained Video Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems face challenges in efficiently transmitting video streams while maintaining detection accuracy, as they require either powerful camera-side processing or high-quality video transmission.
Innovation Solution
The proposed image processing assembly employs a hybrid processing method, where the video stream is highly compressed for visual control, and detection information is encoded and transmitted separately. This allows for reduced data transmission and enables more accurate detection on the client side using higher computational powers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If video streams are transmitted with high quality for accurate detection, then detection accuracy is improved, but data transmission volume increases
Solution Approach 1:
The system segments the video processing into two parts: the camera side performs initial detection to identify regions of interest, and only these specific regions are transmitted in high quality to the client side for verification. This segmentation allows the system to maintain detection accuracy for critical areas while significantly reducing overall data transmission volume compared to transmitting the entire video stream in high quality.
2Measurement precision
If detection is performed on the camera side with high accuracy, then detection precision is improved, but device complexity increases
Solution Approach 1:
The camera side performs partial detection action by identifying only regions of interest rather than performing complete verification on all detected objects. This partial action reduces the computational burden and device complexity on the camera side, while the client side performs the remaining verification tasks, achieving a balance between detection precision and device complexity.
3Measurement precision
If the camera performs all detection tasks with high computational power, then detection accuracy is improved, but energy consumption increases
Solution Approach 1:
The system extracts and transmits only the essential detection information (regions of interest) from the camera side to the client side, rather than performing all computationally intensive detection tasks on the camera. This extraction approach reduces the energy consumption on the camera while maintaining detection accuracy through client-side verification using higher computational power.
Data Source
AI summary
The invention relates to an image processing assembly 1 having a transmission device 2, comprising: a camera interface 5 for acquiring image data 9 from a camera 6; a video encoder 14 for encoding the image data 9 and generating encoded image data; a transmitting detector device 8 for detecting image features 10 with a first detection certainty in feature image areas 11 of the image data 9 and a cropping device 12 for cropping the feature image areas 11, wherein the image features 10 with the feature image areas 11 comprise content data, a content encoder 13 for encoding the content data and generating encoded content data, a transmission interface 6 for transmitting the encoded image data and the encoded content data; with a receiving device 3, comprising: a receiving interface 7 for receiving the encoded image data and the encoded content data; a video decoder device 15 for decoding the encoded image data and generating decoded image data; a content decoder device 16 for decoding the encoded content data and for generating decoded content data, and at least one receiving detector device 17 for detecting and/or verifying the image features with a second detection certainty based on the decoded content data.