Non-maximum suppression acceleration method, system and equipment for target detection
A non-maximum value suppression and target detection technology, applied in the field of image recognition, can solve the problems that affect the convergence speed of the non-maximum value suppression algorithm and the time-consuming non-maximum value suppression algorithm, so as to improve the screening speed and improve the grouping speed. The effect of precision
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0043] In order to select an optimal bounding box from multiple candidate bounding boxes, a non-maximum suppression algorithm (NMS algorithm for short) is usually used to "suppress" bounding boxes with low confidence. Because the non-maximum value suppression algorithm uses the method of searching for the local maximum value to suppress the maximum value, that is, when suppressing or deleting similar candidate bounding boxes, it is necessary to compare one by one until the best bounding box is found. When the number of candidate bounding boxes in the feature map is large, the algorithm takes more time, which affects the overall convergence speed of the algorithm. Based on this, this embodiment provides a non-maximum suppression acceleration method for target detection, which improves the convergence speed of the non-maximum suppression algorithm and reduces the time consumption of the algorithm.
[0044] like figure 1 shown, acceleration methods include:
[0045] obtaining a...
Embodiment 2
[0082] like Figure 4 As shown, the second embodiment provides a non-maximum suppression acceleration system for target detection on the basis of the first embodiment, including:
[0083] a feature extraction unit, configured to obtain a target image containing a detection target, perform feature extraction on the target image, and obtain a feature map, wherein the feature map includes a plurality of first candidate frames;
[0084] a clustering and grouping unit, configured to use a clustering algorithm to perform clustering processing on the center points of a plurality of first candidate frames of the feature map to obtain multiple groups of first candidate frames;
[0085] a deduplication unit, used for simultaneously removing redundant first candidate frames in multiple groups of first candidate frames by using a non-maximum suppression algorithm to obtain a second candidate frame;
[0086] The processing unit is configured to perform regression processing on the second ...
Embodiment 3
[0089] like Figure 5 As shown, this embodiment provides an electronic device, such as Figure 5 As shown, it includes: a processor 310, a communication interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communication interface 320 and the memory 330 communicate with each other through the communication bus 340; the memory 330 is used for storing A computer program; the processor 310 is used for executing the program stored in the memory 330, and realizes the following steps: acquiring a target image including a detection target, performing feature extraction on the target image, and obtaining a feature map, wherein the feature map includes Multiple first candidate frames; use clustering algorithm to cluster the center points of multiple first candidate frames in the feature map to obtain multiple sets of first candidate frames; use non-maximum suppression algorithm to remove multiple sets of first candidate frames simultaneously The redun...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More