Home Publications

Tharun Anand

I'm a Senior Research Engineer at Samsung Research India working on XR and 3D vision with a focus on controllable 3D generation.

Before, I completed my Dual Degree (B.Tech + M.Tech) in Electrical Engineering at IIT Madras, where I worked with Prof. Kaushik Mitra and Prof Pravin Nair in the Computer Vision Lab.

I've spent summers at Aalto University and Intel Labs India where I worked with Prof. Arno Solin in 2025, and Dr. Venkat Natarajan in 2024.

Email / Github / LinkedIn / Scholar

Tharun Anand

Publications

Go to [Publications ] Tab to view my works.

News

  • 📰 [2025/10] Samsung Galaxy XR released — congratulations to our XR team!
  • 📝 [2025/09] Paper submitted to ICASSP 2026 - Self Attention Decomposition for Training Free Diffusion Editing
  • 🔬 [2025/06] Completed ASCII Research Fellowship at Aalto University with Dr. Arno Solin (March - June 2025)
  • 💼 [2025/06] Joined Samsung Research India as Senior Research Engineer (Immersive XR and 3D Vision team)
  • 🎓 [2025/05] Graduated with Dual Degree (B.Tech + M.Tech) from IIT Madras
  • 📄 [2025/03] Paper accepted at CVPR-W 2025 - Detecting Localized Deepfake Manipulations Using Action Unit-Guided Video Representations
  • 🔬 [2025/03] Started ASCII Research Fellowship at Aalto University with Dr. Arno Solin (Interpretability in Diffusion Models)
  • 📄 [2025/01] Paper accepted at ICLR-W 2025 - Learning Self-Supervised Style Representations for Detecting AI-Generated Faces
  • 📄 [2024/12] Paper accepted at WACV 2025 - IP-Face-Diff: Identity Preserving Facial Video Editing with Diffusion
  • 💼 [2024/06] Completed Research Internship at Intel Labs (Robotics & Computer Vision with Dr. Venkat Natarajan)
  • 💼 [2024/03] Started Research Internship at Intel Labs (Bangalore)
  • 🥉 [2023/XX] Bronze Medal at ISRO Moon Super-Resolution National Hackathon
  • 🔬 [2023/05] Started Masters Thesis at Computational Imaging Lab, IIT Madras (Diffusion Models for Generative Video Modelling)

Research Interests

I am passionate about computer vision and its interdisciplinary applications. I have innovated in generative diffusion models for Video synthesis and editing. I also have expertise in deepfake/anamoly detection and have devoloped detection models for detecting challenging deepfakes with fewer forgery traces from recent generative models. My long term goal is to advance the potential of generative models to solve important tasks in multidisciplinary areas while simultaneously focusing on developing generalizable models to detect and prevent their misuse.

  • Diffusion Models: Developing efficient Text-to-Video Diffusion Models with context preservation.
  • DeeepFake Detection: Creating AI solutions for detecting photorealistic deepfake media by analyzing subtle watermarks left by generative models .
  • Audio-Video Understanding: Learning joint representations between Audio-Video modalities to solve downstream tasks such as lip-syncing , emotion recognition and talking face generation.
[Detailed Resume ]