Calcium imaging is a key method in neuroscience for investigating patterns of neuronal activity in vivo . Still, existing algorithms to detect and extract activity signals from calcium-imaging movies have major shortcomings. We introduce the HNCcorr algorithm for cell identification in calcium-imaging datasets that addresses these shortcomings. HNCcorr relies on the combinatorial clustering problem HNC (Hochbaum’s Normalized Cut), which is similar to the Normalized Cut problem of Shi and Malik, a well known problem in image segmentation. HNC identifies cells as coherent clusters of pixels that are highly distinct from the remaining pixels. HNCcorr guarantees a globally optimal solution to the underlying optimization problem as well as minimal dependence on initialization techniques. HNCcorr also uses a new method, called “similarity squared”, for measuring similarity between pixels in calcium-imaging movies. The effectiveness of HNCcorr is demonstrated by its top performance on the Neurofinder cell identification benchmark. We believe HNCcorr is an important addition to the toolbox for analysis of calcium-imaging movies.

No abstract

No abstract

No abstract

Incremental heuristic search algorithms are a class of heuristic search algorithms applicable to the problem of goal-directed navigation. D* and D*Lite are among the most well-known algorithms for this problem. Recently, two new algorithms have been shown to outperform D*Lite in relevant benchmarks: Multi-Path Adaptive A* (MPAA*) and D*ExtraLite. Existing empirical evaluations, unfortunately, do not allow to obtain meaningful conclusions regarding the strengths and weaknesses of these algorithms. Indeed, in the paper introducing D*ExtraLite, it is shown that D*Lite outperforms MPAA* in benchmarks in which the authors of MPAA* claim superiority over D*Lite. The existence of published contradictory data unfortunately does not allow practitioners to make decisions over which algorithm to use given a specific application. In this paper, we analyze two factors that significantly influence the performance of MPAA*, explaining why it is possible to obtain very different results depending on such factors. We identify a configuration of MPAA* which, in the majority of the benchmark problems we use, exhibits superior performance when compared to both D*Lite and D*ExtraLite. We conclude that MPAA* should be the algorithm of choice in goal-directed navigation scenarios in which the heuristic is accurate, whereas D*ExtraLite should be preferred when the heuristic is inaccurate.