*4.4. Effectiveness of Multi-Scale Balanced Sampling on Negative Samples*

To evaluate the effectiveness of multi-scale balanced sampling, comparison of negative samples' distribution on different scales and IOU between MB-RPN, FPN and Libra RCNN sampling method on DOTA dataset, the result are shown in Figure 8, where *b*<sup>1</sup> ∼ *b*<sup>8</sup> denote total amount of hard\_small to easy\_big negative samples. Most of the samples are Easy\_small in FPN method, the Libra RCNN alleviate this situation significantly but the majority is still easy samples(*b*1, *b*<sup>3</sup> and *b*6), especially the easy\_small samples, the multi-scale sampling further improved samples distribution situation. The matching results are also visualized in Figure 9, compare to FPN and Libra RCNN, the matching results has been effectively improved.

**Figure 8.** Distribution of multi-scale balanced,FPN and Libra RCNN sampling method.

**Figure 9.** Example of Multi-scale Balanced Sampling Results of FPN, Libra RCNN and Multi-scale Balanced.
