Layer-Wise Query Selection to Eliminate Redundant Queries in DETR

Recent advancements in the detection Transformer model have demonstrated remarkable accuracy in real-time object detection using an end-to-end approach. DETR leverages the concept of object queries, which act as “questions” to determine the presence and location of objects. However, the excessive nu...

Full description

Saved in:
Bibliographic Details
Main Authors: Seok-Jin Hong, Chan-Young Choi, Sang-Woong Lee
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/14/7686
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849714296323309568
author Seok-Jin Hong
Chan-Young Choi
Sang-Woong Lee
author_facet Seok-Jin Hong
Chan-Young Choi
Sang-Woong Lee
author_sort Seok-Jin Hong
collection DOAJ
description Recent advancements in the detection Transformer model have demonstrated remarkable accuracy in real-time object detection using an end-to-end approach. DETR leverages the concept of object queries, which act as “questions” to determine the presence and location of objects. However, the excessive number of object queries significantly increases computational complexity, leading to higher training and inference times, greater memory consumption, and increased costs. To address this issue, this paper introduces the selected query detection Transformer, a novel approach that optimizes the selection of object queries in the decoder. By progressively filtering and reducing unnecessary queries, the selected query detection Transformer maintains model performance while significantly reducing computational overhead, resulting in faster training and inference times.
format Article
id doaj-art-4f1fe1823ef045be815ac39b66623f50
institution DOAJ
issn 2076-3417
language English
publishDate 2025-07-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-4f1fe1823ef045be815ac39b66623f502025-08-20T03:13:44ZengMDPI AGApplied Sciences2076-34172025-07-011514768610.3390/app15147686Layer-Wise Query Selection to Eliminate Redundant Queries in DETRSeok-Jin Hong0Chan-Young Choi1Sang-Woong Lee2Pattern Recognition and Machine Learning Laboratory, School of Computing, Gachon University, Seungnam 13306, Republic of KoreaPattern Recognition and Machine Learning Laboratory, School of Computing, Gachon University, Seungnam 13306, Republic of KoreaPattern Recognition and Machine Learning Laboratory, School of Computing, Gachon University, Seungnam 13306, Republic of KoreaRecent advancements in the detection Transformer model have demonstrated remarkable accuracy in real-time object detection using an end-to-end approach. DETR leverages the concept of object queries, which act as “questions” to determine the presence and location of objects. However, the excessive number of object queries significantly increases computational complexity, leading to higher training and inference times, greater memory consumption, and increased costs. To address this issue, this paper introduces the selected query detection Transformer, a novel approach that optimizes the selection of object queries in the decoder. By progressively filtering and reducing unnecessary queries, the selected query detection Transformer maintains model performance while significantly reducing computational overhead, resulting in faster training and inference times.https://www.mdpi.com/2076-3417/15/14/7686object detectionquery selectiondetection Transformer
spellingShingle Seok-Jin Hong
Chan-Young Choi
Sang-Woong Lee
Layer-Wise Query Selection to Eliminate Redundant Queries in DETR
Applied Sciences
object detection
query selection
detection Transformer
title Layer-Wise Query Selection to Eliminate Redundant Queries in DETR
title_full Layer-Wise Query Selection to Eliminate Redundant Queries in DETR
title_fullStr Layer-Wise Query Selection to Eliminate Redundant Queries in DETR
title_full_unstemmed Layer-Wise Query Selection to Eliminate Redundant Queries in DETR
title_short Layer-Wise Query Selection to Eliminate Redundant Queries in DETR
title_sort layer wise query selection to eliminate redundant queries in detr
topic object detection
query selection
detection Transformer
url https://www.mdpi.com/2076-3417/15/14/7686
work_keys_str_mv AT seokjinhong layerwisequeryselectiontoeliminateredundantqueriesindetr
AT chanyoungchoi layerwisequeryselectiontoeliminateredundantqueriesindetr
AT sangwoonglee layerwisequeryselectiontoeliminateredundantqueriesindetr