Application of Pattern Spectra and Convolutional Neural Networks to the Analysis of Simulated Cherenkov Telescope Array Data

Abstract

The Cherenkov Telescope Array (CTA) will be the next generation gamma-ray observatory and will be the major global instrument for very-high-energy astronomy over the next decade, offering 5 − 10 × better flux sensitivity than current generation gamma-ray telescopes. Each telescope will provide a snapshot of gamma-ray induced particle showers by capturing the induced Cherenkov emission at ground level. The simulation of such events provides images that can be used as training data for convolutional neural networks (CNNs) to determine the energy of the initial gamma rays. Compared to other state-of-the-art algorithms, analyses based on CNNs promise to further enhance the performance to be achieved by CTA. Pattern spectra are commonly used tools for image classification and provide the distributions of the shapes and sizes of various objects comprising an image. The use of relatively shallow CNNs on pattern spectra would automatically select relevant combinations of features within an image, taking advantage of the 2D nature of pattern spectra. In this work, we generate pattern spectra from simulated gamma-ray events instead of using the raw images themselves in order to train our CNN for energy reconstruction. This is different from other relevant learning and feature selection methods that have been tried in the past. Thereby, we aim to obtain a significantly faster and less computationally intensive algorithm, with minimal loss of performance.

Event reconstruction using pattern spectra and convolutional neural networks for the Cherenkov Telescope Array

“…in [14]. The final results, including the performance on the Alpha Configuration of CTA, are discussed in [30].…”

“…Other reconstruction methods are based on semi-analytical or Gaussian photosphere shower models, such as model analysis [9] and 3D model analysis [10]. Lately, convolutional neural networks (CNNs) have been applied to IACT data [11][12][13][14][15] motivated by their success in other image classification and regression tasks [16]. However, one of the main drawbacks of this method is that the training of CNNs is computationally very expensive [17].…”

“…Figure2: Top-row: CTA image example of a 1.9 TeV gamma-ray event captured by eight SSTs with detected features highlighted in red/orange. Bottom-row: pattern spectrum extracted from the CTA image with pixel corresponding to the detected features marked in red (adapted from[12]). …”

Deep learning techniques for Imaging Air Cherenkov Telescopes

“…With advances in computational power, more recent studies have attempted to leverage advanced ML techniques such as Deep Neural Networks (DNNs) for background rejection (gamma-hadron separation) at IACTs [31][32][33][34][35][36][37][38][39][40]. Some of these studies have also tried to leverage DNNs to reconstruct other shower properties as well, such as the energy and angle of the primary particle [41,42]. DNNs are a type of Artificial Neural Network that possess many layers which allow them to extract complex features of a raw input data set in a highly efficient manner.…”