REVIEW 6 cited by
NMS Strikes Back
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Detection Transformer (DETR) directly transforms queries to unique objects by using one-to-one bipartite matching during training and enables end-to-end object detection. Recently, these models have surpassed traditional detectors on COCO with undeniable elegance. However, they differ from traditional detectors in multiple designs, including model architecture and training schedules, and thus the effectiveness of one-to-one matching is not fully understood. In this work, we conduct a strict comparison between the one-to-one Hungarian matching in DETRs and the one-to-many label assignments in traditional detectors with non-maximum supervision (NMS). Surprisingly, we observe one-to-many assignments with NMS consistently outperform standard one-to-one matching under the same setting, with a significant gain of up to 2.5 mAP. Our detector that trains Deformable-DETR with traditional IoU-based label assignment achieved 50.2 COCO mAP within 12 epochs (1x schedule) with ResNet50 backbone, outperforming all existing traditional or transformer-based detectors in this setting. On multiple datasets, schedules, and architectures, we consistently show bipartite matching is unnecessary for performant detection transformers. Furthermore, we attribute the success of detection transformers to their expressive transformer architecture. Code is available at https://github.com/jozhang97/DETA.
Forward citations
Cited by 6 Pith papers
-
Adversarial Attention Perturbations for Large Object Detection Transformers
AFOG, a learnable-attention white-box attack, collapses object detection mAP across twelve transformer detectors and three CNN detectors using small, visually imperceptible perturbations.
-
SNAP: A Benchmark for Testing the Effects of Capture Conditions on Fundamental Vision Tasks
A new benchmark shows that camera capture settings and lighting systematically change the performance of image classifiers, object detectors, and VQA models, and that common vision datasets are biased toward narrow ex...
-
Robust and Label-Efficient Deep Waste Detection
An ensemble-based soft pseudo-labeling pipeline improves waste detection on the ZeroWaste dataset, beating fully supervised training with the same labeled images.
-
DS-Det: Single-Query Paradigm and Attention Disentangled Learning for Flexible Object Detection
A flexible-query transformer detector that separates cross-attention localization from self-attention deduplication reports consistent accuracy gains over DINO across five backbones.
-
World Model-Based End-to-End Scene Generation for Accident Anticipation in Autonomous Driving
Adding world-model-generated synthetic driving videos to training data, together with a dynamic graph and dilated temporal model, improves accident anticipation accuracy and lead time on multiple benchmarks.
-
Contrasting Cognitive Styles in Vision-Language Models: Holistic Attention in Japanese Versus Analytical Focus in English
Japanese-prompted vision-language models produce more background-first captions than English-prompted ones, but the effect is confounded by the evaluator and by language grammar.
Discussion (0). Continue with ORCID to comment.