Novel tracking approach based on fully-unsupervised disentanglement of the geometrical factors of variation
share
Efficient tracking algorithm is a crucial part of particle tracking detectors. While big work was done in designing plethora of various algorithms, they usually require tedious tuning for each use case. (Weakly) supervised Machine Learning-based approaches can leverage the actual raw data for maximal performance. Yet in realistic scenarios sufficient high-quality labeled data is not available. While sometimes training can be performed on simulated data, often appropriate simulation of detector noise is impossible, compromising this approach. Here we propose a novel fully unsupervised approach to track reconstruction. The introduced model for learning to disentangle the factors of variation in a geometrically meaningful way employs geometrical space invariances. We train it through constraints on the equivariance between the image space and the latent representation in a Deep Convolutional Autoencoder. Using experimental results on synthetic data we show requirement of the variety of the space transformations for meaningful disentanglement of factors of variation. We also demonstrate performance of our model on real data from tracking detectors.
READ FULL TEXT