Published projects | Michele Antonazzi

Robotics

Instance-Guided Unsupervised Domain Adaptation for Robotic Semantic Segmentation

In this work, we propose an unsupervised domain adaptation approach for semantic segmentation in indoor robotics. Our method extends the multi-view consistency by adding an instance refinement step to propagate semantic categories inside object instances.

Privacy-Preserving Robotic Perception for Object Detection in Curious Cloud Robotics

This paper addresses the privacy issues of cloud-based object detection for mobile robots. Our approach involves co-training an encoder-decoder architecture to retain only task-specific features while obfuscating sensitive information, leveraging a novel mechanism of weak loss with proposal selection.

R2SNet: Scalable Domain Adaptation for Object Detection in Cloud-Based Robotic Ecosystems via Proposal Refinement

R2SNet aims to mitigate performance degradation from domain shifts by adapting the object detection process to the robot's target environment, focusing on relabeling, rescoring, and suppression of bounding-box proposals.

Multi-robot rendezvous in communication-restricted unknown environments via backtracking and semantic frontier-based exploration

We adapt a standard frontier-based exploration technique to integrate exploration and rendezvous into a unified strategy, with a mechanism that allows robots to re-visit previously explored regions thus enhancing rendezvous opportunities.

Development and Adaptation of Robotic Vision in the Real-World: the Challenge of Door Detection

We demonstrate with an extensive methodological evidence how the integration of simulation frameworks and domain adaptation techniques can create an effective development pipeline for constructing door detectors specifically tailored for mobile service robots operating in the real world.