Parallelizing image analysis algorithms:: ANET solution and performances

Several hard problems have to be addressed in order to parallelize image analysis algorithms. Indeed, at the region level, these algorithms handle irregular (and sometimes strongly dynamic) data-structures. Moreover, they often lead to unbalanced amount of computations, which is quite impossible to forseen offline. This paper focus on the parallelization of the ANET image analysis programming environment. Thanks to graphs related data structures and efficient computing primitives, ANET allow rapid image algorithms prototyping [1]. But in return, these primitives are difficult to parallelize. We present a solution for powerful implicit parallelization of the ANET environment, whitout any change in the application programming interface. The ANET API is summarizing and illustrating with some examples. Several parallelizations experimentations are reported. The solution we propose is detailled, and results are given on complete image analysis applications. ANET appears as a powerful environment, both for its expressiveness that allow rapid prototyping than for its implicit parallelization that allow good computation time.