MDB 2026

hosted by

Ethical Video Analytics in the Wild: Methods, Datasets and Benchmarks

August 31th, 2026

EVENT STARTS IN

-
Days
-
Hours
-
Minutes
-
Seconds

Call for Papers

The growing presence of cameras worldwide and the complexity of real-world environments demand robust and responsible video analytics solutions that operate effectively “in the wild”. This workshop, MDB2026, aims to bridge the gap between controlled research settings and unconstrained real-world scenarios. We welcome contributions that address the core challenges of video understanding in such environments, including (but not limited to) object detection, multi-target and multi-camera tracking, re-identification, activity recognition, anomaly detection, and environmental monitoring, all under the demanding conditions of operating in real scenarios in the wild.

In addition to technical aspects, MDB2026 encourages discussions on fairness, transparency, privacy preservation, and accountability in video analytics systems. The workshop emphasizes the importance of methods, datasets, and benchmarks that foster reproducibility, scalability and ethical compliance: in fact, MDB2026 aims to advance the state of the art in video analytics through innovative algorithms, comprehensive evaluation protocols, and open resources, also taking into account the computational requirements. Case studies and demonstrations from real-world domains are strongly encouraged.

MDB2026 aims to provide a dedicated forum for researchers, practitioners, and industry representatives to present and discuss innovative contributions in this emerging field, encouraging interdisciplinary collaboration and knowledge exchange.

Research Topics

Core Technologies & Behavioral Analysis
  • Object Detection and Recognition
  • Multi-Target and Multi-Camera Tracking
  • Person and Object Re-Identification
  • Activity and Behavior Analysis
  • Anomaly Detection and Event Forecasting
  • Video Understanding in Unconstrained Environments
Trustworthy AI, Data & Real-World Impact
  • Datasets and Benchmarks for Video Analytics
  • Domain Adaptation and Generalization
  • Environmental and Situational Monitoring
  • Case Studies and Real-World Applications
  • Privacy-Preserving Video Analytics
  • Fairness and Bias Mitigation in Video Understanding

Submission Guidelines

MDB 2026 invites submissions of full papers, presenting mature and complete research with substantial contributions and validated results.

Submissions must adhere to the IEEE double-column format and are limited to six pages, including figures, tables, and references. All accepted papers presented at MDB 2026 will be included in the conference proceedings and published in the IEEE Xplore digital library.

The Microsoft CMT service is used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support. The paper review process will be double-blind to ensure fairness and impartiality.

Call for Contest

AUTOMATIC INCIDENT DETECTION CHALLENGE (AIDC26)

FIRST EDITION

AIDC26 aimed at collecting and evaluating advanced methods for identifying accidents on the roads in real-time videos recorded in real situations.

The dataset that we will use for the challenge is the MIVIA-AID dataset, comprising more than 2300 videos, with more than six hours of footage and over ninety minutes of annotated accident intervals. In more details, we incorporate about 1200 clips from seven different state-of-the-art datasets, after filtering to retain only fixed-camera footage, and complement them with an additional about 900 videos acquired directly from real surveillance systems. Such videos were sourced from openly accessible online platforms, such as YouTube, various social networks, and publicly available municipal webcams, ensuring a diverse and representative set of real surveillance scenarios.

CRIME DETECTION CHALLENGE (CDC26)

FIRST EDITION

CDC26 aimed at proposing and benchmarking robust algorithms for detecting crime, violent or aggressive behaviors in real-time videos obtained in real-world scenarios.

The dataset provided to participants consists of a training set of 1,900 videos, 950 containing crimes and 950 normal videos, derived from the widely used UCF-Crime dataset. This dataset has been further augmented with temporal annotations specifying the start and end times of anomalous events. The crimes are categorized into 13 classes: Abuse, Arrest, Arson, Assault, Burglary, Explosion, Fighting, Road Accident, Robbery, Shooting, Shoplifting, Stealing, and Vandalism. In addition, a private test set, which will not be released to participants, comprises 270 previously unseen video clips (135 containing crimes and 135 normal videos). These clips were collected from real-world unconstrained scenarios “in the wild.”

Important dates

PAPER SUBMISSION DEADLINE

April 25th, 2026

--
days
--:--:--
CONTEST PAPER SUBMISSION DEADLINE

May 5th, 2026

--
days
--:--:--
DECISION NOTIFICATION DEADLINE

May 25th, 2026

--
days
--:--:--
CAMERA-READY DEADLINE

June 1st, 2026

--
days
--:--:--

Organizers

VINCENZO CARLETTI

Department of Information Engineering, Electrical Engineering, and Applied Mathematics (DIEM)
University of Salerno, Italy

vcarletti@unisa.it

ANDREA CAVALLARO

EPFL, Switzerland

andrea.cavallaro@epfl.ch

ANTONIO GRECO

Department of Information Engineering, Electrical Engineering, and Applied Mathematics (DIEM)
University of Salerno, Italy

agreco@unisa.it

ALESSIA SAGGESE

Department of Information Engineering, Electrical Engineering, and Applied Mathematics (DIEM)
University of Salerno, Italy

asaggese@unisa.it

BRUNO VENTO

Department of Electrical Engineering and Information Technology (DIETI)
University of Naples - Federico II, Italy

bruno.vento@unina.it

Contact Us

For any questions or information regarding the workshop, please contact us at:

mdb2026@unisa.it