Skip to content

AUC / Prof. Yasser Gadallah

Autonomous sensing research

One Research Assistant role, three connected efforts: build the acquisition platform, collect context-rich data, and investigate crossing prediction.

01 / Acquire

ARGUS

30 Hz platform; 50+ hours of recordings.

02 / Document

EUPB

Cairo stereo dataset captured at 15 Hz.

03 / Predict

Crossing intent

Attention-based fusion of 2D and 3D features.

01 / Research systems

RESEARCH SYSTEM / CONCEPTUAL30 Hz

Sensors → sessions → research data

ARGUS

A field-deployable platform that turns stereo cameras into a real-time, recordable, browser-controlled sensing station for urban research.

~30 HzView case study

02 / Dataset research

RESEARCH SYSTEM / CONCEPTUAL15 Hz

Cairo / urban pedestrian behavior

EUPB dataset

Stereo imagery, depth and motion context from Cairo’s urban traffic.

15 HzView case study

03 / Machine learning

RESEARCH SYSTEM / CONCEPTUAL

Complementary features, one fusion pipeline

Pedestrian crossing prediction

Combining image features with 3D pedestrian key-points in a PyTorch prediction pipeline.

PyTorchView case study

Publications & reports

Research in progress

Under review: a manuscript on pedestrian action prediction, with Mazen as second author, at the International Journal of Computer Vision. ARGUS, EUPB and crossing prediction each target a research publication; no accepted journal papers are claimed here.

TET report section: AUCP_TET-RS-001, 4 September 2026, sole-authored by Mazen. Supporting internal report: AUCP_TET-TR-001.

TET public abstract: April 2026, co-authored with Youssef I. Aref, Youssef I. Mohammed and Janna Lelah.