Research Paper#Anomaly Detection, Optical TPC, Autoencoders, Data Reduction🔬 ResearchAnalyzed: Jan 3, 2026 17:16
Fast ROI Triggering with Autoencoders in Optical TPCs
Published:Dec 30, 2025 15:28
•1 min read
•ArXiv
Analysis
This paper presents a novel approach for real-time data selection in optical Time Projection Chambers (TPCs), a crucial technology for rare-event searches. The core innovation lies in using an unsupervised, reconstruction-based anomaly detection strategy with convolutional autoencoders trained on pedestal images. This method allows for efficient identification of particle-induced structures and extraction of Regions of Interest (ROIs), significantly reducing the data volume while preserving signal integrity. The study's focus on the impact of training objective design and its demonstration of high signal retention and area reduction are particularly noteworthy. The approach is detector-agnostic and provides a transparent baseline for online data reduction.
Key Takeaways
- •Introduces an unsupervised, reconstruction-based anomaly detection method for fast ROI extraction in optical TPCs.
- •Employs convolutional autoencoders trained on pedestal images to learn detector noise morphology.
- •Achieves high signal retention and significant image area reduction.
- •Demonstrates the importance of training objective design for effective anomaly detection.
- •Provides a detector-agnostic baseline for online data reduction.
Reference
“The best configuration retains (93.0 +/- 0.2)% of reconstructed signal intensity while discarding (97.8 +/- 0.1)% of the image area, with an inference time of approximately 25 ms per frame on a consumer GPU.”