Fig 1.
(a) C2F module (b) Efficient modulation module.
Fig 2.
Structure of Bilinear interpolation upsample, input feature map (X), upsampled feature map (X’).
Fig 3.
Structure of DySample original grid (G).
Fig 4.
The structure of YOLO-ED.
Table 1.
LUNG-PET-CT-DX diagnosis dataset partition.
Table 2.
LUNA16 diagnosis dataset partition.
Table 3.
Training parameters.
Fig 5.
Prediction box and ground truth box.
Table 4.
Results of the integrating different attention modules on Lung-PET-CT-DX.
Table 5.
Results of the improved upsample part on LUNG-PET-CT-DX.
Table 6.
Results of the improved attention and upsample part on LUNG-PET-CT-DX.
Table 7.
Ablation experimental results on LUNG-PET-CT-DX.
Table 8.
Performance comparison of the different models on LUNG-PET-CT-DX.
Table 9.
Performance comparison of the different models on LUNA16.
Fig 6.
Lung cancer classification results on LUNG-PET-CT-DX.
Fig 7.
Precision change curve on LUNG-PET-CT-DX.
Fig 8.
Precision change curve on LUNA16.
Fig 9.
YOLO-ED detection effect on LUNG-PET-CT-DX.
Fig 10.
YOLO-ED detection effect on LUNA16.