**Input:** Two image A (source image) and B (target image); |

Dense local descriptors F_{A} of A and F_{B} of B. |

**Output:** Corresponding field **T**. |

1: Divide A to the non-overlapping cells of size *w*×*w* and obtain the BoW histograms for each cell *i*. |

2: Divide B to overlapping cells of size *w*×*w* with step size *s* and obtain the BoW histograms of each cell *j*. |

3: Determine the most similar cell centered at (*x*_{B}, *y*_{B}) in B for each cell *c*_{A} in A. |

4: Set **T**(*x*, *y*)=(*x*_{B}, *y*_{B}) ((*x*, *y*)∈*c*_{A}) for each cell *c*_{A} in A and random permute of **T**. |

5: Update **T** by field propagation as PatchMatch algorithm using the local descriptors F_{A} and F_{B}. |

6: Update **T** by locally randomized search using the local descriptors F_{A} and F_{B}. |

7: Repeat steps 5 and 6 a fixed number of times or until convergence. |