Pith. sign in

REVIEW 9 cited by

HazyDet: Open-Source Benchmark for Drone-View Object Detection with Depth-Cues in Hazy Scenes

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

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

arxiv 2409.19833 v2 pith:SP2AYZR7 submitted 2024-09-30 cs.CV

Changfeng Feng , Zhenyuan Chen , Xiang Li , Chunping Wang , Jian Yang , Ming-Ming Cheng , Yimian Dai , Qiang Fu This is my paper · ORCID
classification cs.CV
keywords hazydetdetectionhazydecodetdomainobjectbenchmarkchallenging
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Object detection from aerial platforms under adverse atmospheric conditions, particularly haze, is paramount for robust drone autonomy. Yet, this domain remains largely underexplored, primarily hindered by the absence of specialized benchmarks. To bridge this gap, we present \textit{HazyDet}, the first, large-scale benchmark specifically designed for drone-view object detection in hazy conditions. Comprising 383,000 real-world instances derived from both naturally hazy captures and synthetically hazed scenes augmented from clear images, HazyDet provides a challenging and realistic testbed for advancing detection algorithms. To address the severe visual degradation induced by haze, we propose the Depth-Conditioned Detector (DeCoDet), a novel architecture that integrates a Depth-Conditioned Kernel to dynamically modulate feature representations based on depth cues. The practical efficacy and robustness of DeCoDet are further enhanced by its training with a Progressive Domain Fine-Tuning (PDFT) strategy to navigate synthetic-to-real domain shifts, and a Scale-Invariant Refurbishment Loss (SIRLoss) to ensure resilient learning from potentially noisy depth annotations. Comprehensive empirical validation on HazyDet substantiates the superiority of our unified DeCoDet framework, which achieves state-of-the-art performance, surpassing the closest competitor by a notable +1.5\% mAP on challenging real-world hazy test scenarios. Our dataset and toolkit are available at https://github.com/GrokCV/HazyDet.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. More Clear, More Flexible, More Precise: A Comprehensive Oriented Object Detection benchmark for UAV

    cs.CV 2025-04 conditional novelty 7.0 of 10

    CODrone is a 10,004-image, 12-class UAV dataset with oriented boxes, and a benchmark showing that current detectors lose 10 to 15 AP points from 30m to 100m altitude and from nadir to 30-degree views.

  2. Advancing MLLM-based UAV Image Understanding and Reasoning: A Benchmark and a Training-Free Multi-Agent System

    cs.CV 2026-08 conditional novelty 6.0 of 10

    A new self-constructed UAV QA benchmark and a training-free multi-agent system that claims state-of-the-art accuracy on it without training.

  3. FRFDet: Efficient UAV Small Object Detection with Symmetric Sampling and Scalable Fusion

    cs.CV 2026-07 accept novelty 6.0 of 10

    FRFDet improves UAV small-object detection with learnable symmetric Inverse Bidirectional Sampling and capacity-dependent Scale-Feature Relationship Cross-Fusion, reaching SOTA among lightweight models.

  4. Predictive Modeling for High Impact Active Learning Classrooms

    physics.ed-ph 2026-03 unverdicted novelty 6.0 of 10

    Four classroom-time variables predict physics concept learning, and classes with 10–20% group worksheets, 20–40% group clickers, and ≥2 student questions per hour show effect sizes above 2.

  5. SNNSIR: A Simple Spiking Neural Network for Stereo Image Restoration

    cs.CV 2025-08 unverdicted novelty 6.0 of 10

    A fully spike-driven network with residual, modulation, and cross-attention modules reports competitive stereo image restoration across rain, raindrop, low-light, and super-resolution tasks.

  6. UAVScenes: A Multi-Modal Dataset for UAVs

    cs.CV 2025-07 conditional novelty 6.0 of 10

    UAVScenes adds frame-wise image and LiDAR semantic labels, reconstructed 6-DoF poses, and 3D maps to 120k frames of the MARS-LVIG dataset, with six benchmark tasks.

  7. A Task-Driven Evaluation of UAV Detection and Tracking under Synthetic Fog

    cs.CV 2026-07 conditional novelty 5.0 of 10

    Fog degrades UAV detection and tracking mainly via missed detections; fog-inclusive training is more robust than test-time dehazing, and restoration quality does not proportionally improve downstream perception.

  8. Teaching in adverse scenes: a statistically feedback-driven threshold and mask adjustment teacher-student framework for object detection in UAV images under adverse scenes

    cs.CV 2025-06 conditional novelty 5.0 of 10

    A dynamic mask-and-threshold teacher-student framework for unsupervised domain adaptation reports mAP50 66.9 on HazyDet and top transfer scores on RDDTS and DroneVehicle.

  9. Fine-Tuning Adversarially-Robust Transformers for Single-Image Dehazing

    cs.CV 2025-04 conditional novelty 5.0 of 10

    Pre-trained single-image dehazing transformers are vulnerable to one-pixel and small L-infinity perturbations, and scale-and-bias or linear-adapter fine-tuning with adversarial training restores much of the robustness...

Pith tools