in computer vision and image processing [28, 7, 27, 2, 6], e.g., denoising [26], enhancement
[23], inpainting [5], segmentation [17], stereo and optical flow computation.
In general, there are two kinds of methods used to design PDEs. For the first kind of
methods, PDEs are written down directly, based on some mathematical understandings on
the properties of the PDEs (e.g., anisotropic diffusion [26], shock filter [23] and curve evolu-
tion based equations [28, 27, 2, 6]). The second kind of methods basically define an energy
functional first, which collects the wish list of the desired properties of the output image,
and then derives the evolution equations by computing the Euler-Lagrange variation of the
energy functional (e.g., chapter 9 of [28]). Both methods require good skills in choosing ap-
propriate functions and predicting the final effect of composing these functions such that the
obtained PDEs roughly meet the goals. In either way, people have to heavily rely on their
intuition, e.g., smoothness of edge contour and surface shading, on the vision task. Such
intuition should easily be quantified and be described using the operators (e.g., gradient and
Laplacian), functions (e.g., quadratic and square root functions) and numbers (e.g., 0.5 and
1) that people are familiar with. As a result, the designed PDEs can only reflect very limited
aspects of a vision task (hence are not robust in handling complex situations in real appli-
cations) and also appear rather artificial. If people do not have enough intuition on a vision
task, they may have difficulty in acquiring effective PDEs. For example, can we have a PDE
(or a PDE system) for object detection (Figure 1) that locates the object region if the object
is present and does not respond if the object is absent? We believe that this is a big challenge
to human intuition because it is hard to describe an object class, which may have significant
variation in shape, texture and pose. Although there has been much work on PDE-based
image segmentation, e.g., [17], the basic philosophy is always to follow the strong edges
of the image and also require the edge contour to be smooth. Without using additional in-
formation to judge the content, the artificial PDEs always output an “object region” for any
non-constant image. In short, current PDE design methods greatly limit the applications
of PDEs to wider and more complex scopes. This motivates us to explore whether we can
acquire PDEs that are not artificial yet more powerful.
2