Object Detection Device Using Segmented Image Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object detection techniques, such as YOLOX, are inefficient in obtaining detection results due to the trade-off between accuracy and time, with higher accuracy requiring longer processing times.
Innovation Solution
The implementation of an object detection device with multiple detection units of varying accuracy, where an input image is divided into partial images and processed by designated units, followed by Non-Maximum Suppression to eliminate overlapping bounding box errors, allowing for faster result generation without sacrificing accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single high-accuracy detection unit is used to process the entire input image, then detection accuracy is improved, but processing time increases significantly
Solution Approach 1:
The input image is divided into multiple partial images corresponding to different partial regions. Each partial image is processed by a detection unit, allowing parallel processing and reducing overall processing time while maintaining detection accuracy through region-specific optimization.
Solution Approach 2:
Different detection units with different accuracy levels are assigned to different partial regions based on their characteristics. High-accuracy detection units are deployed to regions requiring precise detection, while lower-accuracy units handle regions where speed is prioritized, optimizing the balance between accuracy and processing time.
2Productivity
If multiple detection units process different partial regions in parallel, then processing time is reduced, but bounding box overlap and detection errors increase
Solution Approach 1:
An NMS (Non-Maximum Suppression) unit is introduced as an intermediary component that receives detection results from multiple detection units. The NMS unit eliminates overlapping bounding boxes by suppressing redundant detections, thereby resolving conflicts between parallel processing outputs and improving overall detection reliability.
3Measurement precision
If all partial images are input to all detection units for comprehensive processing, then detection coverage is improved, but computational complexity and processing time increase
Solution Approach 1:
Each detection unit is designated to process specific partial images corresponding to its optimal detection regions. This localized assignment reduces unnecessary computational operations while ensuring each region is processed by the most suitable detection unit, balancing coverage and complexity.
Solution Approach 2:
The image and detection units are segmented such that each detection unit receives only the partial images it is designated to process. This segmentation eliminates redundant processing of all images by all units, reducing computational complexity while maintaining comprehensive coverage through coordinated processing of all partial regions.
Data Source
AI summary
The object detection device includes two or more detection means that perform detection operation of deriving bounding boxes of a detection target object and scores indicating reliability of the bounding boxes, based on a given image. Accuracy of the detection operation of each detection means is different. The division means divides an input image to obtain partial images corresponding to partial regions indicated by region setting values, when a pair of the region setting values that indicate the partial regions in the input image and designation information designating the detection means to be input destination for partial images corresponding to the partial regions is given. The partial image input means inputs the individual partial images to designated detection means. Each NMS means eliminates bounding box overlap in overlapping bounding boxes output from corresponding detection means.