Leonardo Peroni

This is the codebase for the publication titled In-Band Quality Notification from Users to ISPs, presenting Youstall, a prototype mechanism that enables users to signal YouTube stall duration to ISPs along the server-client path. Users install this software to notify ISPs of QoE issues, enabling ISPs to improve connection quality without costly and often inefficient encrypted packet analysis. The software includes a user agent that detects stalls on YouTube and triggers a specific packet pattern by interacting with the YouTube interface, and an ISP agent that detects this pattern for possible responsive action.

This is the codebase for the publication titled Quality of Experience in Video Streaming: Status Quo, Pitfalls, and Guidelines. Through an in-depth analysis of two large databases—Waterloo-IV¹ and iQoE²—as well as a comprehensive literature review, it identifies common pitfalls in QoE practices and models in video streaming. These issues are classified into test conduct, model building, and model application, with proposed solutions for improvement. This codebase specifically presents the database analyses.

This codebase accompanies the publication Empowerment of Atypical Viewers via Low-Effort Personalized Modeling of Video Streaming Quality. It introduces iQoE, a novel method for building personalized QoE models by engaging individual users in a brief series of subjective assessments. The method leverages an active learning paradigm with the RIGS sampler and XSVR modeler. This codebase includes the code to replicate all analyses from the paper, the newly generated dataset, and the iQoE method itself.

This codebase includes the implementation for this portfolio showcase website. It uses the popular static site generator Jekyll and is built on the minimal-light theme¹, which itself is based on the minimal template².

This repository contains code for the analysis of the Waterloo-IV database, demonstrating how individual QoE (Quality of Experience) perception significantly differs from group QoE perception. Since QoE models in the literature are constructed based on group perception, the results of this analysis suggest that these models could greatly benefit from personalization. This analysis represents the preliminary stage for the publication Empowerment of Atypical Viewers via Low-Effort Personalized Modeling of Video Streaming Quality.