RT-DETR
PulseAugur coverage of RT-DETR — every cluster mentioning RT-DETR across labs, papers, and developer communities, ranked by signal.
3 天有情绪数据
-
DFIR-DETR 通过细化频域特征改进小目标检测
研究人员开发了 DFIR-DETR，一种用于复杂视觉场景中小目标检测的新方法。该方法解决了现有神经网络设计中的根本性局限性，例如均匀的注意力分布以及空间卷积对高频细节的抑制。DFIR-DETR 专门针对上采样特征中的范数漂移和关键边缘组件的丢失等问题。该模型在 NEU-DET 和 VisDrone 数据集上展示了显著的性能提升，在参数量和计算成本相对较低的情况下实现了高 mAP50 分数。
-
New method enhances VLM document layout understanding
Researchers have developed a new method to improve how Vision-Language Models (VLMs) understand document layouts, particularly for documents with structures not seen during training. The approach pre-resolves layout inf…
-
New PaQ-RT-DETR model improves multi-class battery detection accuracy
Researchers have developed a new method called PaQ-RT-DETR for detecting multiple types of batteries, aiming to improve accuracy and efficiency in applications like electronic waste recycling and quality control. They e…
-
New BEM module suppresses false positives in real-time camera detection
Researchers have developed a new training-free module called Background Embedding Memory (BEM) designed to improve the accuracy of object detectors in real-world scenarios. BEM works by estimating background embeddings …
-
无人机除草检测模型在准确性和速度之间取得平衡，适用于边缘设备
研究人员开发了一个框架，用于在资源受限的无人机上部署除草检测模型，以实现位点特异性管理。该研究评估了包括YOLO和RT-DETR变体在内的各种目标检测模型，这些模型部署在Jetson Orin Nano和Jetson AGX Xavier等不同的边缘设备上。结果表明，在检测准确性和计算效率之间存在权衡，高容量模型实现了更好的准确性但推理时间较慢。轻量级模型提供了实时性能，而RT-DETRv2-R50-M和YOLOv11s在实际无人机应…