Please use this identifier to cite or link to this item: https://hdl.handle.net/1959.11/61991
Title: Confluence: A Robust Non-IoU Alternative to Non-Maxima Suppression in Object Detection
Contributor(s): Shepley, Andrew J  (author)orcid ; Falzon, Gregory  (author)orcid ; Kwan, Paul  (author); Brankovic, Ljiljana  (author)orcid 
Publication Date: 2023-10
DOI: 10.1109/TPAMI.2023.3273210
Handle Link: https://hdl.handle.net/1959.11/61991
Abstract: 

Confluence is a novel non-Intersection over Union (IoU) alternative to Non-Maxima Suppression (NMS) in bounding box post-processing in object detection. It overcomes the inherent limitations of IoU-based NMS variants to provide a more stable, consistent predictor of bounding box clustering by using a normalized Manhattan Distance inspired proximity metric to represent bounding box clustering. Unlike Greedy and Soft NMS, it does not rely solely on classification confidence scores to select optimal bounding boxes, instead selecting the box which is closest to every other box within a given cluster and removing highly confluent neighboring boxes. Confluence is experimentally validated on the MS COCO and CrowdHuman benchmarks, improving Average Precision by 0.2--2.7% and 1--3.8% respectively and Average Recall by 1.3--9.3 and 2.4--7.3% when compared against Greedy and Soft-NMS variants. Quantitative results are supported by extensive qualitative analysis and threshold sensitivity analysis experiments support the conclusion that Confluence is more robust than NMS variants. Confluence represents a paradigm shift in bounding box processing, with potential to replace IoU in bounding box regression processes.

Publication Type: Journal Article
Source of Publication: IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(10), p. 11561-11574
Publisher: Institute of Electrical and Electronics Engineers
Place of Publication: United State of America
ISSN: 1939-3539
0162-8828
2160-9292
Fields of Research (FoR) 2020: 460304 Computer vision
Socio-Economic Objective (SEO) 2020: 220499 Information systems, technologies and services not elsewhere classified
Peer Reviewed: Yes
HERDC Category Description: C1 Refereed Article in a Scholarly Journal
Appears in Collections:Journal Article
School of Science and Technology

Show full item record

SCOPUSTM   
Citations

18
checked on Dec 14, 2024
Google Media

Google ScholarTM

Check

Altmetric


Items in Research UNE are protected by copyright, with all rights reserved, unless otherwise indicated.