This website is to demonstrate a machine learning tool, MorphoMatcher, which is to match the neuronal images of Fruitfly neurons between two databases, FlyCircuit and FlyEM databases. This is a considerable challenge due to the drastic difference in their resolutions of neuronal images. We propose MorphoMatcher to address this problem by leveraging morphological comparisons, which could effectively reduce the influence of resolution differences. Using more than 1219 FC-EM pairs confirmed by human experts as benchmark training data, we develop MorphoMatcher by combining Case-Based Reasoning and Machine Learning techniques. Here we provide the matching results for neuroscience research and model prediction for new finding neurons. More data and matching results will be updated in the near future. Similar method can be also extended to the identification of neuron images with other databases.
Brain Research Center, National Tsing Hua Unversity
AIFR Group, National Tsing Hua Unversity