Automated tracking and analysis of axonal transport using combined filtering methods

Abstract

This paper describes an automated technique to accurately record and quantitatively describe the movements of mitochondria along the axon from time-lapse images acquired from a microfluidic culture platform using fluorescence microscopy. For this, kymograph is generated using a vessel tracing technique from a consolidated image obtained by combining all frames of input video. In order to enhance ridge structures in kymograph, two filtering techniques are applied, the matched and the vessel enhancement filters. A point tracking algorithm is then applied to detect mitochondria trajectories. All processing steps for filtering and trajectory detection are performed automatically. In order to assess the performance of the proposed technique quantitatively, experiments have been performed using a set of real image sequences. All results were compared with ground truth data obtained by tracing object trajectories manually. Our method provides good accuracy and robustness sufficient for practical use and it is expected to facilitate the quantitative analysis of axonal transport experiments required for neuroscience research.