Efficient Non-Maximum Suppression

@article{Neubeck2006EfficientNS,
  title={Efficient Non-Maximum Suppression},
  author={A. Neubeck and L. Gool},
  journal={18th International Conference on Pattern Recognition (ICPR'06)},
  year={2006},
  volume={3},
  pages={850-855}
}
  • A. Neubeck, L. Gool
  • Published 2006
  • Computer Science
  • 18th International Conference on Pattern Recognition (ICPR'06)
In this work we scrutinize a low level computer vision task - non-maximum suppression (NMS) - which is a crucial preprocessing step in many computer vision applications. Especially in real time scenarios, efficient algorithms for such preprocessing algorithms, which operate on the full image resolution, are important. In the case of NMS, it seems that merely the straightforward implementation or slight improvements are known. We show that these are far from being optimal, and derive several… Expand
A Convnet for Non-maximum Suppression
TLDR
This work proposes a convnet designed to perform non-maximum suppression of a given set of detections, and overcomes the intrinsic limitations of greedy NMS, obtaining better recall and precision. Expand
Robust Low Complexity Corner Detector
TLDR
This paper redesigned Harris and KLT algorithms to reduce their complexity in each stage of the algorithm: Gaussian derivative, cornerness response, and non-maximum suppression (NMS), achieving a complexity reduction by a factor of 9.8. Expand
A search based approach to non maximum suppression in face detection
TLDR
Results show how the application of this new non-maximum suppression approach to a simple face detector boosts its performance to state of the art results. Expand
Lococo: low complexity corner detector
TLDR
The complexity of Harris and KLT corner detectors are reduced by using the box kernel, the integral image and efficient non-maximum suppression, achieving complexity reduction by a factor of 8. Expand
Improvement of Non-Maximum Suppression in RGB-D Object Detection
TLDR
An RGB-D object detection network model based on the YOLO v3 framework is designed, and using level-by-level metaphase fusion on the RGB and depth information, an improved NMS algorithm which fuses depth characteristics is proposed which can not only detect more overlapping objects but also achieve a better object localization accuracy. Expand
Fast Linear Feature Detection Using Multiple Directional Non-Maximum Suppression
TLDR
A new algorithm for linear feature detection using multiple directional non-maximum suppression is developed that is very fast compared with methods in the literature and very good results have been obtained. Expand
Non-maximum Suppression Using Fewer than Two Comparisons per Pixel
TLDR
Two solutions that use fewer than 2 comparisons per pixel with little memory overhead are presented that are considerably faster than current best methods in the literature when applied to feature point detection. Expand
A new fast efficient non-maximum suppression algorithm based on image segmentation
TLDR
A new fast efficient non-maximum suppression algorithm based on image segmentation and border analysis is proposed, which allowed to increase the accuracy of local extremum extraction in comparison with known non- maximum suppression algorithms and reduce the computational complexity and the use of RAM in relation with the morphological algorithms. Expand
Sample Distillation for Object Detection and Image Classification
TLDR
It is shown that distillation can be used to improve the traditional bootstrapping approach to object detection and using the result of a distillation process instead of a random sub-set taken uniformly in the original sample set improves performance signicantly. Expand
Improved robustness and efficiency for automatic visual site monitoring
TLDR
A novel implementation of a pedestrian detector that runs on commodity graphics hardware and is up to 76 faster than the authors' CPU-only implementation is produced, and it is shown how one can combine a variety of detection and tracking techniques to robustly handled event detection scenarios such as theft and left-luggage detection. Expand
...
1
2
3
4
5
...

References

SHOWING 1-5 OF 5 REFERENCES
Distinctive Image Features from Scale-Invariant Keypoints
This paper presents a method for extracting distinctive invariant features from images that can be used to perform reliable matching between different views of an object or scene. The features are ...
Computing 2-D Min, Median, and Max Filters
TLDR
Fast algorithms for computing min, median, max, or any other order statistic filter transforms are described and a logarithmic time per pixel lower bound for the computation of the median filter is shown. Expand
Scale & Affine Invariant Interest Point Detectors
TLDR
A comparative evaluation of different detectors is presented and it is shown that the proposed approach for detecting interest points invariant to scale and affine transformations provides better results than existing methods. Expand
Distinctive image features from scale-invar iant keypoints.IJCV
  • 2004
Matching Widely Separated Views Based on Affine Invariant Regions
TLDR
To increase the robustness of the system, two semi-local constraints on combinations of region correspondences are derived (one geometric, the other photometric) allow to test the consistency of correspondences and hence to reject falsely matched regions. Expand