Skip to main content

Table 2 Performance of the proposed vs existing models on VisDrone 2019 VID dataset

From: MSRP-TODNet: a multi-scale reinforced region wise analyser for tiny object detection

Models

Performance metrics

Accuracy

Precision

Recall

F1-Score

AUC-ROC

STDNet [11]

92.86%

93.00%

92.34%

92.79%

0.655

Ex-FPN [12]

94.89%

93.07%

93.40%

94.98%

0.716

QueryDet [13]

95.86%

95.99%

95.84%

95.97%

0.826

YOLO-FIRI [14]

97.15%

97.65%

97.34%

95.26%

0.927

Proposed (MSRP-TODNet)

99.23%

99.94%

99.08%

99.99%

0.997

Back to article page

BMC Research Notes

ISSN: 1756-0500

Contact us