Scholarlink

<p>Object detection is a pivotal task in computer vision, with applications in numerous areas including surveillance, autonomous driving, and image analysis. Traditional methods for object detection required handcrafted feature extraction and the use of classifiers to label objects within images. With the advent of deep learning, these processes have been significantly automated, leading to advancements in both accuracy and efficiency of object detection systems. In this work, we introduce a novel architecture that incorporates a Dense Prior module and Focal Self-attention mechanism into the DEtection TRansformer (DETR) framework. This Transformed DETR model is designed to direct greater focus to regions of interest in images, thereby improving detection performance. We demonstrate the effectiveness of our approach on the COCO val2017 dataset using a ResNet-50 backbone, achieving a Mean Average Precision (mAP) of 53.8%, which indicates a considerable improvement over existing methods including the baseline DETR (42.0% mAP). Our architecture captures and focuses on the key regions within an image, resulting in a more powerful object detection model.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Transformed DETR leveraging dense prior and focal attention for enhanced visual recognition

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月30日

查看全文
  • Hakim Nasaoui,
  • Insaf Bellamine,
  • Hassan Silkan

摘要

Object detection is a pivotal task in computer vision, with applications in numerous areas including surveillance, autonomous driving, and image analysis. Traditional methods for object detection required handcrafted feature extraction and the use of classifiers to label objects within images. With the advent of deep learning, these processes have been significantly automated, leading to advancements in both accuracy and efficiency of object detection systems. In this work, we introduce a novel architecture that incorporates a Dense Prior module and Focal Self-attention mechanism into the DEtection TRansformer (DETR) framework. This Transformed DETR model is designed to direct greater focus to regions of interest in images, thereby improving detection performance. We demonstrate the effectiveness of our approach on the COCO val2017 dataset using a ResNet-50 backbone, achieving a Mean Average Precision (mAP) of 53.8%, which indicates a considerable improvement over existing methods including the baseline DETR (42.0% mAP). Our architecture captures and focuses on the key regions within an image, resulting in a more powerful object detection model.