A Small Object Detection Algorithm Based on Modulated Deformable Convolution and Large Kernel Convolution

<div>Ghost model. Conv represents convolution, <svg height="12.3916pt" id="M1" style="vertical-align:-3.42948pt" version="1.1" viewbox="-0.0498162 -8.96212 10.2016 12.3916" width="10.2016pt" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink"><g transform="matrix(.013,0,0,-0.013,0,0)"><path d="M542 242C542 369 469 429 379 468C370 472 372 496 376 513L416 681L400 691L344 648L219 51C167 69 113 128 113 217C113 324 166 387 266 428L252 459C124 418 23 336 23 202C23 95 90 13 205 -8L182 -132C165 -222 191 -254 209 -261L215 -258L270 -7C415 -4 542 98 542 242ZM469 232C469 126 403 33 279 39L355 409C399 394 469 331 469 232Z"></path></g><g transform="matrix(.0091,0,0,-0.0091,7.02,3.132)"><path d="M250 606C250 634 233 656 203 656C168 656 146 618 146 593C146 564 169 545 192 545C227 545 250 573 250 606ZM227 95L212 119C187 98 152 71 135 71C129 71 128 78 134 102L207 373C219 418 217 451 194 451C165 451 92 411 30 351L44 326C77 353 106 371 114 371C124 371 121 357 117 341L55 97C32 5 46 -12 70 -12C108 -12 191 51 227 95Z"></path></g></svg> represents linear operation, and identity represents identity mapping.</div>

Computational Intelligence and Neuroscience

fig2

Figure 2

Figure 2: A Small Object Detection Algorithm Based on Modulated Deformable Convolution and Large Kernel Convolution 