Skip to main content
Fig. 3 | Brain Informatics

Fig. 3

From: DeepNeuron: an open deep learning toolbox for neuron tracing

Fig. 3

Axon detection results on two challenging datasets. (a) An example of a 3D image stack of mouse neuron imaged with bright-field microscopy (also see Fig. 2a). (b) An example of a 3D stack (shown as cropped) from a whole mouse brain imaged with fMOST; the pixel resolution is 0.3 um × 0.3 um × 1 um. Red dots indicate detected axonal signals. The two false positive example patches shown in (a) and (b) could be eliminated with more training samples. Those false negative patches (also as shown in (a) and (b)) with very weak signals at the center could be further identified by increasing the amount of weak signal foreground samples in the training set

Back to article page