Efficient Non-Maximum Suppression

By A. Neubeck, Luc Van Gool. 2006

1,872 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Electrical and Electronic Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

In Electrical and Electronic Engineering

Citations in the corpus

Voting for Voting in Online Point Cloud Object Detection (2015, methodology): “VI. NON-MAXIMUM SUPPRESSION To remove duplicate detections, we apply a non-maximum suppression procedure analogous to the standard techniques commonly applied in Computer Vision [2, 6, 13].”
Learning Human Pose Estimation Features with Convolutional Networks (2013, methodology): “For datasets containing examples with multiple persons (known a priori), we use non-maximal suppression [29] to ﬁnd multiple local maxima across the ﬁltered response-maps from each scale, and we then take the top n most likely joint candidates from each person in the scene.”
End-to-End Text Recognition with Hybrid HMM Maxout Models (2013, methodology): “We follow this pipeline by doing a non-max suppression (NMS) (Neubeck & Van Gool, 2006) on word boxes that overlap by 30% of the area of their bounding box according to the visual cost cost v from the word recognition module.”
Drone-Based RGB-Infrared Cross-Modality Vehicle Detection Via Uncertainty-Aware Learning (2020, methodology): “Additionally, 110 we further design an illumination-aware (IA) non-maximum 111 suppression (NMS) strategy to improve the integration of 112 the detection results of different modalities.”
Speeded-Up Robust Features (SURF) (2008, methodology): “Specifically, we use a fast variant introduced by Neubeck and Van Gool [33].”
Pose Estimation for Augmented Reality: A Hands-On Survey (2016, methodology): “According to gravity direction, a captured image is ﬁrst rectiﬁed and orientations from both texture and gravity are used to enhance the distinctiveness of descriptors.”
A Forest Fire Detection System Based on Ensemble Learning (2021, background): “Average Precision (AP) AP 0.5 AP at IoU = 0.5”
Occluded Prohibited Items Detection: An X-ray Security Inspection Benchmark and De-occlusion Attention Module (2020, methodology): “cilitates the next detector task. More recently, [4, 12, 13, 33, 39] solved the problem of traditional methods relying heavily on the anchor box, with the only post-processing non-maximum suppression [19]. In this work, we make performance comparisons for our method DOAM with selected approaches fro”
Deep Matching Prior Network: Toward Tighter Multi-oriented Text Detection (2017, background): “…kinds of circumstances: i) redundant information may reduce the reliability of detected conﬁdence [18] and make subsequent recognition harder [40]; ii) marginal text may not be localized completely; iii) when using non-maximum suppression [21], unnecessary overlap may eliminate true prediction.”
DETRs with Hybrid Matching (2022, background): “…transformers [57] to 2 D object detection, more and more follow-up works [44, 13, 10, 62] have built various advanced extensions based on DETR because it removes the need for many hand-designed components like non-maximum suppression [47] or initial anchor boxes generation [16, 51, 32, 35].”
Measuring the Objectness of Image Windows (2011, background): “We run nonmaximum suppression on this 4D score space using the efficient technique of Neubeck and Van Gool [43], resulting in a set of local maxima windows Wsmax.”
Slim-neck by GSConv: A better design paradigm of detector architectures for autonomous vehicles (2022, background): “The former may be more controllable for designers and users, but the NMS [23] must be used to filter out the prediction box with the highest IoU threshold score.”
TPH-YOLOv5: Improved YOLOv5 Based on Transformer Prediction Head for Object Detection on Drone-captured Scenarios (2021, methodology): “There are three different methods to ensemble boxes from different object detection models: Non-maximum suppression (NMS) [36], Soft-NMS [53], weighted boxes fusion (WBF) [43].”
State of the Art in Defect Detection Based on Machine Vision (2021, methodology): “The traditional object detection strategies and algorithms include Viola-Jones [183], HOG + SVM, non-maximum suppression (NMS) [184], the deformable part model (DPM) [185], selective search [186, 187], and edge boxes [188].”
CAT: Cross Attention in Vision Transformer (2021, methodology): “AdamW[76] optimizer, 1x schedule, and NMS[80] are employed.”