DT = VL_IMDISTTF(I) computes the distance transform of image I. The distance transform is defined as

    DT(x,y) = min_{u,v} I(u,v) + (u - x)^2 + (v - y)^2.


[D, INDEXES] = VL_IMDISTTF(...) returns a matrix INDEXES that contains for each pixel (x,y) the index of the pixel (u,v) which is the minimizer of the distance transform objective.

VL_IMDSITTF(I, PARAM) uses differenet parameters for the transform:

   DT(u,v) = min_{u,v} I(u,v) + PARAM(1) (u - x - PARAM(2))^2
+ PARAM(3) (v - y - PARAM(4))^2


The function uses the fast algorithm from [1].

REFERENCES

[1] P. F. Felzenszwalb and D. P. Huttenlocher. Distance transforms of sampled functions. Technical report, Cornell University, 2004.