Niloufar Alipour Talemi

Computer Vision | Machine Learning | Vision-Language Models

Welcome to my homepage!

I am currently pursuing my Ph.D. in the IS-WiN Lab at Clemson University under the supervision of Dr. Fatemeh Afghah. My research focuses on advancing generalizable vision-language models , emphasizing the CLIP framework for vision-based applications. Before joining Clemson University, I worked as a research assistant at West Virginia University (January 2022 – May 2024), working on biometric applications such as facial attribute recognition and identity verification.

It has been an honor to serve as a reviewer for over 60 papers, contributing to prestigious conferences such as CVPR, AAAI, WACV, ICCV, and ICIP, as well as esteemed journals including Computers in Biology and Medicine, ACM Transactions on Computing for Healthcare, Knowledge-Based Systems and IEEE Access.

News

[2025/02] Our recent work, FreqDebias, has been accepted to CVPR 2025.
[2025/01] Awarded the WACV 2025 DEI Scholarship.
[2024/11] Successfully passed Qualifying Exam.
[2024/10] Two papers have been accepted to WACV 2025.
[2024/04] Our recent work CATFace is accepted by IEEE Transactions on Biometrics, Behavior, and Identity Science.
[2023/09] Our new method on Morph Attack Detection has been accepted by IJCB 2023.
[2023/06] AAFace is accepted to IEEE ICIP 2023 as an oral presentation.

Publications

	Style-Pro: Style-Guided Prompt Learning for Generalizable Vision-Language Models. NA Talemi, H Kashiani, F Afghah IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025. arxiv / poster We propose a style-guided prompt learning framework to enhance generalization in Vision-Language models.
	ROADS: Robust Prompt-driven Multi-Class Anomaly Detection under Domain Shift. H Kashiani, NA Talemi , F Afghah IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025. arxiv / poster We propose a robust multi-class anomaly detection framework with a class-aware prompt integration mechanism to mitigate inter-class interference and a domain adapter to handle domain shifts.
	CATFace: Cross-Attribute-Guided Transformer With Self-Attention Distillation for Low-Quality Face Recognition. NA Talemi, H Kashiani, NM Nasrabadi IEEE Transactions on Biometrics, Behavior, and Identity Science, 2024. arxiv / IEEE / bibtex We propose a novel multi-branch neural network with cross-attribute-guided fusion and self-attention distillation, improving face recognition in low-quality images using soft biometric attributes.
	AAFACE: Attribute-aware Attentional Network for Face Recognition. NA Talemi , H Kashiani, and 4 more authors IEEE International Conference on Image Processing (ICIP), 2023. arxiv / IEEE / bibtex / poster We present a multi-branch network using attribute-aware integration to enhance face recognition through soft biometric prediction.
	Towards Generalizable Morph Attack Detection with Consistency Regularization. H Kashiani, NA Talemi, NM Nasrabadi IEEE International Joint Conference on Biometrics (IJCB), 2023. arxiv / IEEE / bibtex / poster / code We propose consistency regularization to enhance the generalization of morph attack detection through morph-wise augmentations to enhance robustness against unseen morph attacks in biometric systems.
	Face Image Quality Vector Assessment for Biometrics Applications. N Najafzadeh, H Kashiani, MSE Saadabadi, NA Talemi, and 2 more authors Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2023. CVF / IEEE / bibtex This paper proposes a multi-task neural network that generates a face quality vector, including nuisance factors, offering improved performance and detailed feedback for face image quality assessment.
	Superpixel-based brain tumor segmentation in MR images using an extended local fuzzy active contour model. NA Talemi, RPR Hasanzadeh Multimedia Tools and Applications, 2021. Springer / bibtex This paper introduces a novel region-based fuzzy active contour model, leveraging curve evolution techniques, for brain tumor segmentation in magnetic resonance images, effectively addressing challenges arising from noise and heterogeneity.