Brian K. S. Isaac-Medina

AI Engineer

About Me

Hi! I am Brian, I am an AI Engineer @ OpenWorks Engineering, currently based in the North East of England. My current work is focused in object detection and tracking optimised for edge devices, including implementation of SOTA techniques as well as evaluating current trends in computer vision applied to different business needs. Additionally, I do research focused in Machine Learning for computer vision, with a focus on anomaly detection and out-of-distribution detection. My former research also includes multi-view geometry, object detection, Vision Transformers and neural radiance fields.

I hold a PhD and an MsC in Computer Science from Durham University and an Engineering degree from the Autonomous University of Yucatan, Mexico, and currently I am a mentor for LatinX in AI. If you have any idea of a project or you want to collaborate, reach out!

News!

Publications

FEVER-OOD: Free Energy Vulnerability Elimination for Robust Out-of-Distribution Detection

Brian K. S. Isaac-Medina,  Mauricio Che,  Yona F. A. Gaus,  Samet Akcay,  Toby P. Breckon
ICCV (to appear), 2025
arXiv  code  bibtex 

Dream-Box: Object-wise Outlier Generation for Out-of-Distribution Detection

Brian K. S. Isaac-Medina,  Toby P. Breckon
LatinX in AI, CVPR Workshop, 2025
arXiv  code 

Semi-supervised Object-Wise Anomaly Detection for Firearm and Firearm Component Detection in X-ray Security Imagery

Yona F. A. Gaus, Brian K. S. Isaac-Medina,  Neelanjan Bhowmik, Yee T. Lam,  Toby P. Breckon
VAND, CVPR Workshop, 2025
paper 

Towards Open-World Object-based Anomaly Detection via Self-Supervised Outlier Synthesis

Brian K. S. Isaac-Medina,  Yona F. A. Gaus,  Neelanjan Bhowmik,  Toby P. Breckon
ECCV, 2024
arXiv  code  bibtex  poster 

On Deep Machine Learning for Multi-view Object Detection and Neural Scene Rendering

Brian K. S. Isaac-Medina
PhD Thesis, 2024

Performance Evaluation of Segment Anything Model with Variational prompting for Application to Non-Visible Spectrum Imagery

Yona F. A. Gaus,  Neelanjan Bhowmik, Brian K. S. Isaac-Medina,  Toby P. Breckon
PBVS (CVPR Workshop), 2024
paper  arXiv  code  bibtex 

Exact-NeRF: An Exploration of a Precise Volumetric Parameterization for Neural Radiance Fields

Brian K. S. Isaac-Medina,  Chris G. Willcocks,  Toby P. Breckon
CVPR, 2023
paper  arXiv  code  presentation  bibtex 

Region-based Appearance and Flow Characteristics for Anomaly Detection in Infrared Surveillance Imagery

Yona F. A. Gaus,  Neelanjan Bhowmik, Brian K. S. Isaac-Medina,  Amir Atapour-Abarghouei,  Hubert P. H. Shum,  Toby P. Breckon
VAND (CVPR Workshop), 2023
paper  demo  presentation  bibtex 

Seeing Through the Data: A Statistical Evaluation of Prohibited Item Detection Benchmark Datasets for X-ray Security Screening

Brian K. S. Isaac-Medina,  Seyma Yucer,  Neelanjan Bhowmik,  Toby P. Breckon
PBVS (CVPR Workshop), 2023
paper  bibtex 

Multi-view Vision Transformers for Object Detection

Brian K. S. Isaac-Medina,  Chris G. Willcocks,  Toby P. Breckon
ICPR, 2022
paper  code  presentation  poster  bibtex 

Cross-modal Image Synthesis in Dual-Energy X-Ray Security Imagery

Brian K. S. Isaac-Medina,  Neelanjan Bhowmik,  Chris G. Willcocks,  Toby P. Breckon
PBVS (CVPR Workshop), 2022
paper  bibtex 

UAV-ReID: A Benchmark on Unmanned Aerial Vehicle Re-identification in Video Imagery

Daniel Organisciak,  Matthew Poyser,  Aishah Alsehaim,  Shanfeng Hu, Brian K. S. Isaac-Medina,  Toby P. Breckon,  Hubert P. H. Shum
CoRR, 2021
paper  arXiv  code  bibtex 

Unmanned aerial vehicle visual detection and tracking using deep neural networks: A performance benchmark

Brian K. S. Isaac-Medina,  Matthew Poyser,  Daniel Organisciak,  Chris G. Willcocks,  Toby P. Breckon,  Hubert P. H. Shum
ICCV Workshop, 2021
paper  arXiv  code  demo  bibtex 

Multi-view Object Detection Using Epipolar Constraints within Cluttered X-ray Security Imagery

Brian K. S. Isaac-Medina,  Chris G. Willcocks,  Toby P. Breckon
ICPR, 2020
paper  code  presentation  poster  bibtex 

Visible to Infrared Transfer Learning as a Paradigm for Accessible Real-time Object Detection and Classification in Infrared Imagery

Yona F. A. Gaus,  Neelanjan Bhowmik, Brian K. S. Isaac-Medina,  Toby P. Breckon
Proc. Conf. Counterterrorism, Crime Fighting, Forensics, and Surveillance Technologies, 2020
paper  demo  bibtex 

Projects

Tracking Drones Across Different Platforms with Machine Vision Security Technology Research Innovation Grants Programme (S-TRIG)

Project Website

Providing the UK government with extra capacity to identify and counter hostile drones.

Education

Durham University

PhD Computer Science

Department of Computer Science, 2019 - 2024

A globally outstanding centre of teaching and research excellence, a collegiate community of extraordinary people, a unique and historic setting – Durham is a university like no other.

I did my PhD in the Department of Computer Science at Durham University from 2019 to 2024. I worked under the supervision of Prof. Toby Breckon and I worked in a project for automatic detection of threat items in X-ray security imagery in airports, focusing specifically in the incorporation of multi-view geometry. I also did some research in NeRFs. I received a scholarship from the Mexican Council of Science and Technology (CONACyT) for this period. I was the president of the Durham Mexican Society for the 2020/21 academic year.

Durham University

MSc Internet Systems and E-Business

Department of Computer Science, 2017 - 2018

During my time in the Masters at Durham University I was able to develop my skills as a software developer and project manager. I also got a CONACyT scholarship for this period.

Universidad Autónoma de Yucatán

BEng (Mechatronics)

Faculty of Engineering, 2011 - 2016

La Facultad de Ingeniería tiene como misión la formación integral de recursos humanos, a nivel profesional y de posgrado, en las áreas de las ingenierías y la contribución al avance científico y tecnológico para el desarrollo sostenible de la sociedad.

I did my BEng (Metchatronics) (Licenciatura under the Mexican education system) at the Universidad Autónoma de Yucatán, Mexico. Although my current research is focused in Computer Vision, my early formation in engineering gave me the skills to develop embedded and robotic systems. My undergrad thesis was developed under the supervision of Dr. Francis Avilés.

A Little More About Me

Apart from my work in computer vision, I also enjoy music! I’ve been part of a few band at Durham where I play the drums and the bass. I played the bass with the CVPR houseband since 2023.

Brian K. S. Isaac-Medina -  - References on request