Particle identification in camera image sensors using computer vision

Abstract

We present a deep learning, computer vision algorithm constructed for the purposes of identifying and classifying charged particles in camera image sensors. We apply our algorithm to data collected by the Distributed Electronic Cosmic-ray Observatory (DECO), a global network of smartphones that monitors camera image sensors for the signatures of cosmic rays and other energetic particles, such as those produced by radioactive decays. The algorithm, whose core component is a convolutional neural network, achieves classification performance comparable to human quality across four distinct DECO event topologies. We apply our model to the entire DECO data set and determine a selection that achieves ≥ 90% purity for all event types. In particular, we estimate a purity of 95% when applied to cosmic-ray muons. The automated classification is run on the public DECO data set in real time in order to provide classified particle interaction images to users of the app and other interested members of the public.

Particle Identification in Smartphone Camera Images Using the Distributed Electronic Cosmic-ray Observatory

“…For the purpose of providing real-time classifications for the events listed in the public DECO database, we seek to maintain a high-purity set of events identified as tracks. After evaluating constant cut-off values of 0.7, 0.8 and 0.9 on the test set, we opted for a probability threshold of 0.8, which yields an event selection with a track efficiency 3 of 80%, and, most importantly, a track purity of 90% [5].…”

“…These events were assigned a correct human label and added to the training data available to the next iteration of the model. The full details of the architecture, data augmentation, and training are presented in [5], but we summarize some notable results below.…”

Recognition of Cosmic Ray Images Obtained from CMOS Sensors Used in Mobile Phones by Approximation of Uncertain Class Assignment with Deep Convolutional Neural Network

“…As artefacts, we understand all images taht cannot be attributed to particles’ passage through the sensor but rather result from the deficiencies of the registration procedure. Our approach is based on the morphological properties of particle tracks rather than their physical interpretation, although some studies [ 1 , 2 , 3 ] associated certain shapes of tracks like spots, wiggles (which we here call worms), etc., with muons, electrons, etc. Unambiguous mapping between track shapes and radiation types, however, requires detailed studies of radiation propagation in a sensor of given geometry.…”

“…Combining these features makes the mobile phones an ideal framework for creating the global network of radiation detectors coupled to central data storage. This idea underpinned several particle detection initiatives like CRAYFIS [ 15 , 16 , 17 , 18 , 19 ], DECO [ 1 , 3 , 20 , 21 ], and CREDO [ 22 , 23 ]. The analysis presented in this paper is based on the CREDO detection data set, as this is currently the largest publicly available data set of particle images obtained with mobile phones.…”

“…From the perspective of motivation, the methods and specific tested architectures in our work are similar to those of [ 1 ] (project DECO). However, there are significant differences in image labeling for the classification purpose between our data set and that from DECO, which has convinced us that it is worth trying a different approach than the one proposed so far.…”