Atif Khurshid

About Me

I am a Computer Scientist working to understand the mechanisms behind human and artificial visual intelligence. I am interested in developing computational models of human visual search to inform the design of more robust, efficient, and explainable AI systems.

Currently, I am a PhD student at Ulm University, where I am associated with the Institute for Neural Information Processing and the Research Training Group on Medical XAI — Knowledge Infusion and Extraction for Explainable Medical AI. My research focuses on computational modeling of the visual search mechanisms used by radiologists during diagnostic image reading.

News

[Sep. 2026] Our paper about learning scene gist is accepted at ECCV 2026 Workshop “Human-inspired Computer Vision”.
[Aug. 2026] I will be presenting our investigations into invariance properties of eye movement models at ECEM 2026.
[May. 2025] I have joined the KEMAI research training group at the University of Ulm as PhD student.
[Apr. 2025] Our paper about the TVFace dataset is accepted to PAA.
[Jul. 2024] Our paper about cephalometric landmark detection is accepted to ESWA.
[Nov. 2023] Our paper about knee osteoarthritis classification is accepted to ICOSST 2023.
[Aug. 2023] I have joined the AI & Data Science department at FAST-NUCES as Lecturer.
[Apr. 2021] Our paper about transfer learning grammar is accepted to ICoDT2 2021.

Resume

2025 - Present	PhD Student	UUlm
2023 - 2024	Lecturer	FAST NUCES
2022 - 2023	Research Assistant	MachVIS
2020 - 2023	MS in Computer Science	NUST-SEECS
2016 - 2019	BSc in Computer Science	CUHK CSE

Services

Workshop Reviewer

Developmental Perspectives on AI (DevAI) at NeurIPS 2026

Publications

Invariance properties of predicted eye movements in visual search models

Atif Khurshid*, Matthias Kohn*, Heiko Neumann

European Conference on Eye Movements (ECEM), 2026.

Slides Code *Equal contribution

Learning Task-Specific Scene Gist Representations from Low-Level Image Statistics

Atif Khurshid, Heiko Neumann

European Conference on Computer Vision Workshops (ECCVW), 2026.

Slides Poster Code

TVFace: Towards Large-scale Unsupervised Face Recognition in Video Streams

Atif Khurshid, Bostan Khan, Muhammad Shahzad & Muhammad Moazam Fraz

Pattern Analysis and Applications (PAA), 2025.

Code BibTex

Two-stage Regression Framework for Automated Cephalometric Landmark Detection Incorporating Semantically Fused Anatomical Features and Multi-head Refinement Loss

Muhammad Anwaar Khalid, Atif Khurshid, Kanwal Zulfiqar, Ulfat Bashir, Muhammad Moazam Fraz

Expert Systems with Applications (ESWA), 2024.

Code BibTex

Automated Deep Learning Based Knee Osteoarthritis Joint Extraction and Classification

Muhammad Sohail Abbas, Sonain Jamil, Atif Khurshid

17th International Conference on Open Source Systems and Technologies (ICOSST), 2023.

BibTex

Transfer Learning Grammar for Multilingual Surface Realisation

Atif Khurshid, Seemab Latif, Rabia Latif

International Conference on Digital Futures and Transformative Technologies (ICoDT2), 2021.

Code BibTex