Algorithm 1 Extract–Append Algorithm

Require: Pretrained semantic segmentation model $\mathit{\Phi }$ Input: Input image containing an object ${\mathit{I}}_{\mathit{o}\mathit{b}\mathit{j}}$, Background image ${\mathit{I}}_{\mathit{b}\mathit{a}\mathit{c}\mathit{k}}$ Output: Create new image ${\mathit{A}}_{\mathit{o}\mathit{b}\mathit{j}}$ 1: ${\mathit{M}}_{\mathit{S}}\leftarrow$ $\mathit{\Phi }\left({\mathit{I}}_{\mathit{o}\mathit{b}\mathit{j}}\right)⊳$ Extract the mask of an object 2: $\widehat{{\mathit{M}}_{\mathit{S}}}=\left\{\begin{array}{c}1, \mathit{o}\mathit{b}\mathit{j}\mathit{e}\mathit{c}\mathit{t} \\ 0,\mathit{b}\mathit{a}\mathit{c}\mathit{k}\mathit{g}\mathit{r}\mathit{o}\mathit{u}\mathit{n}\mathit{d}\end{array}\right. \leftarrow$ Binarization of ${\mathit{M}}_{\mathit{S}}$ 3: for each iteration do 4: ${\mathit{E}}_{\mathit{o}\mathit{b}\mathit{j}}\leftarrow {\mathit{I}}_{\mathit{o}\mathit{b}\mathit{j}} \mathit{⨀} \widehat{{\mathit{M}}_{\mathit{S}}} ⊳$ Extract an object from ${\mathit{I}}_{\mathit{o}\mathit{b}\mathit{j}}$ 5: ${\mathit{C}}_{\mathit{b}\mathit{a}\mathit{c}\mathit{k}}\leftarrow {\mathit{I}}_{\mathit{o}\mathit{b}\mathit{j}} \mathit{⨀} \left|1-\widehat{{\mathit{M}}_{\mathit{S}}}\right|⊳$ Making room for object insertion in ${\mathit{I}}_{\mathit{b}\mathit{a}\mathit{c}\mathit{k}}$ 6: ${\mathit{A}}_{\mathit{o}\mathit{b}\mathit{j}} \leftarrow {\mathit{E}}_{\mathit{o}\mathit{b}\mathit{j}}+{\mathit{C}}_{\mathit{b}\mathit{a}\mathit{c}\mathit{k}}$ 7: end for