NET: a new framework for the vectorization and examination of network data

Table 1 Comparison of node number N, network length L, mean edge weight \(\bar {R}\), ratio of smallest to largest weight r=min(R)/max(R) and pixel-wise difference D of the plotted networks

σ _N	σ _L	\(\sigma _{\bar {R}}\)	σ _r	D
9.0±8.6%	2.6±4.6%	25.6±10.0%	24.1±20.1%	0.1±0.1%

For each value, the respective observable was calculated in the known and the re-extracted network and the absolute difference normalized by the value in the known network was calculated. The errors σ _N, σ _L, \(\sigma _{\bar {R}}\), σ _r and D shown in the table are the mean error and its standard deviation for 389 automatically extracted and re-extracted networks by NET from images of Drosophila tracheoles. Error distributions are heavily skewed towards small values, errors cannot be negative

ISSN: 1751-0473