Lossy Event Compression: From Distortion Metrics to Adaptive Codec Design

September 2, 2026

Event cameras generate asynchronous, sparse data streams with microsecond temporal resolution, but in moderate-to-high motion scenes they can produce hundreds of millions of events per second, creating significant bandwidth and storage challenges. Lossy compression is therefore essential, yet reliable event stream distortion metrics capable of predicting compression-induced degradation at the task level are still lacking, forcing codec optimization to rely on expensive task-specific evaluations. To address this, this dissertation introduces two fundamentally different event compression pipelines: i) an aggregation-based pipeline that converts the event stream into polarity-based histogram frames coded with JPEG~2000, and ii) a frame-free point cloud-based pipeline that codes events natively as 3D points using G-PCC. Both are assessed within a task-driven evaluation framework that relates event stream distortion to downstream performance across four tasks under a reference-relative protocol: video reconstruction, object detection, optical flow estimation, and delay-sensitive asynchronous feature tracking. Five classification-based distortion metrics are applied to event compression for the first time and benchmarked against existing metrics; results demonstrate that they reliably predict compression-induced task degradation across different coding frameworks. Exploiting these results, this dissertation develops the Mode-Based Aggregation Codec (MAC), which combines multiple coding modes with complementary rate-distortion characteristics and uses an IoU-based distortion measure within an adaptive rate-distortion mode-selection mechanism. The evaluation over the JPEG~XE development sequences shows that the adaptive MAC improves upon fixed-mode coding for most sequences. Overall, event stream distortion assessment supports both the design and the rate-distortion optimization of lossy event coding solutions.

Authors in the community:

• 

  Z

  Zahra Rezaee

  ist1112031

  Affiliations and sources

  IST

  Instituto Superior Técnico

Supervisors of this institution:

• 

  J

  João Miguel Duarte Ascenso

  ist31104

electrical-engineering-electronic-engineering-information-engineering - Electrical engineering, electronic engineering, information engineering

eng - English

Instituto Superior Técnico