<p>DETR (DEtection TRansformer) is a CV model for object detection that replaces traditional complex methods with a Transformer architecture, and has achieved significant improvement over previous methods, particularly in handling small and medium-sized objects. However, the attention mechanism-based detection framework of DETR exhibits limitations in small and medium-sized object detection. It struggles to extract fine-grained details of small and medium-sized objects from low-resolution features, and its computational complexity increases significantly with the input scale, thereby constraining real-time detection efficiency. To address these limitations, we introduce the Cross Feature Attention (XFA) mechanism and propose XFCOS (XFA-based with FCOS), a novel object detection model built upon it. XFA simplifies the attention mechanism’s computational process and reduces complexity through L2 normalization and two one-dimensional convolutions applied in different directions. This design reduces the computational complexity from quadratic to linear while preserving spatial context awareness. XFCOS enhances the original TSP-FCOS (Transformer-based Set Prediction with FCOS) model by integrating XFA into the transformer encoder, creating a CNN-ViT hybrid architecture, significantly reducing computational costs without sacrificing accuracy. Extensive experiments demonstrate that XFCOS achieves state-of-the-art performance while addressing DETR’s convergence and efficiency limitations. On Pascal VOC 2007, XFCOS attains 54.7 AP and 60.7 AP<InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40747_2025_1904_Article_IEq1.gif" Format="GIF" Height="10" Rendition="HTML" Resolution="72" Type="Linedraw" Width="13" /> </InlineMediaObject> <EquationSource Format="TEX">\(_\textrm{75}\)</EquationSource> <EquationSource Format="MATHML"><math> <mmultiscripts> <mrow /> <mtext>75</mtext> <mrow /> </mmultiscripts> </math></EquationSource> </InlineEquation> - surpassing DETR by 4.5 AP and 7.9 AP<InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40747_2025_1904_Article_IEq1.gif" Format="GIF" Height="10" Rendition="HTML" Resolution="72" Type="Linedraw" Width="13" /> </InlineMediaObject> <EquationSource Format="TEX">\(_\textrm{75}\)</EquationSource> <EquationSource Format="MATHML"><math> <mmultiscripts> <mrow /> <mtext>75</mtext> <mrow /> </mmultiscripts> </math></EquationSource> </InlineEquation> respectively, establishing new benchmarks among ResNet-50-based detectors. The model shows particular strength in small object detection, achieving 24.0 AP<InlineEquation ID="IEq3"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40747_2025_1904_Article_IEq3.gif" Format="GIF" Height="11" Rendition="HTML" Resolution="72" Type="Linedraw" Width="8" /> </InlineMediaObject> <EquationSource Format="TEX">\(_\textrm{S}\)</EquationSource> <EquationSource Format="MATHML"><math> <mmultiscripts> <mrow /> <mtext>S</mtext> <mrow /> </mmultiscripts> </math></EquationSource> </InlineEquation> and 43.9 AP<InlineEquation ID="IEq4"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40747_2025_1904_Article_IEq4.gif" Format="GIF" Height="11" Rendition="HTML" Resolution="72" Type="Linedraw" Width="12" /> </InlineMediaObject> <EquationSource Format="TEX">\(_\textrm{M}\)</EquationSource> <EquationSource Format="MATHML"><math> <mmultiscripts> <mrow /> <mtext>M</mtext> <mrow /> </mmultiscripts> </math></EquationSource> </InlineEquation> on COCO 2017, representing 3.3 AP<InlineEquation ID="IEq5"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40747_2025_1904_Article_IEq3.gif" Format="GIF" Height="11" Rendition="HTML" Resolution="72" Type="Linedraw" Width="8" /> </InlineMediaObject> <EquationSource Format="TEX">\(_\textrm{S}\)</EquationSource> <EquationSource Format="MATHML"><math> <mmultiscripts> <mrow /> <mtext>S</mtext> <mrow /> </mmultiscripts> </math></EquationSource> </InlineEquation> and 3.8 AP<InlineEquation ID="IEq6"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40747_2025_1904_Article_IEq6.gif" Format="GIF" Height="11" Rendition="HTML" Resolution="72" Type="Linedraw" Width="12" /> </InlineMediaObject> <EquationSource Format="TEX">\(_{\textrm{M}}\)</EquationSource> <EquationSource Format="MATHML"><math> <mmultiscripts> <mrow /> <mtext>M</mtext> <mrow /> </mmultiscripts> </math></EquationSource> </InlineEquation> improvements over DETR. Through computational optimization, XFCOS reduces encoder FLOPs to 13.5G, representing a 17.2% decrease compared to TSP-FCOS’s 16.3G, while cutting activation memory from 285.78 to 264.64M, a reduction of 7.4%. This significantly enhances computational efficiency.</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

A lightweight mechanism for vision-transformer-based object detection

  • Yanming Ye,
  • Qiang Sun,
  • Kailong Cheng,
  • Xingfa Shen,
  • Dongjing Wang

摘要

DETR (DEtection TRansformer) is a CV model for object detection that replaces traditional complex methods with a Transformer architecture, and has achieved significant improvement over previous methods, particularly in handling small and medium-sized objects. However, the attention mechanism-based detection framework of DETR exhibits limitations in small and medium-sized object detection. It struggles to extract fine-grained details of small and medium-sized objects from low-resolution features, and its computational complexity increases significantly with the input scale, thereby constraining real-time detection efficiency. To address these limitations, we introduce the Cross Feature Attention (XFA) mechanism and propose XFCOS (XFA-based with FCOS), a novel object detection model built upon it. XFA simplifies the attention mechanism’s computational process and reduces complexity through L2 normalization and two one-dimensional convolutions applied in different directions. This design reduces the computational complexity from quadratic to linear while preserving spatial context awareness. XFCOS enhances the original TSP-FCOS (Transformer-based Set Prediction with FCOS) model by integrating XFA into the transformer encoder, creating a CNN-ViT hybrid architecture, significantly reducing computational costs without sacrificing accuracy. Extensive experiments demonstrate that XFCOS achieves state-of-the-art performance while addressing DETR’s convergence and efficiency limitations. On Pascal VOC 2007, XFCOS attains 54.7 AP and 60.7 AP \(_\textrm{75}\) 75 - surpassing DETR by 4.5 AP and 7.9 AP \(_\textrm{75}\) 75 respectively, establishing new benchmarks among ResNet-50-based detectors. The model shows particular strength in small object detection, achieving 24.0 AP \(_\textrm{S}\) S and 43.9 AP \(_\textrm{M}\) M on COCO 2017, representing 3.3 AP \(_\textrm{S}\) S and 3.8 AP \(_{\textrm{M}}\) M improvements over DETR. Through computational optimization, XFCOS reduces encoder FLOPs to 13.5G, representing a 17.2% decrease compared to TSP-FCOS’s 16.3G, while cutting activation memory from 285.78 to 264.64M, a reduction of 7.4%. This significantly enhances computational efficiency.