Andrei Palaev

About Me

I am a Computer Vision engineer working on perception systems and generative modelling. My work spans detection and classification, segmentation of deformable objects, and controllable diffusion-based image editing. I am particularly interested in statistical and generative modelling.

Research Interests

• Generative and statistical modelling
• Image editing
• 3D generation/reconstruction

Publications

LLM-guided Instance-level Image Manipulation with Diffusion U-Net Cross-Attention Maps BMVC 2024

Instance-level object manipulation using LLM-guided attention control in diffusion models.

Paper Code BibTeX

@inproceedings{Palaev_2024_BMVC,
  title={LLM-guided Instance-level Image Manipulation with Diffusion U-Net Cross-Attention Maps},
  author={Andrey Palaev and Adil Khan and Syed M Ahsan Kazmi},
  booktitle = {35th British Machine Vision Conference 2024, {BMVC} 2024, Glasgow, UK, November 25-28, 2024},
  publisher = {BMVA},
  url={https://papers.bmvc2024.org/0457.pdf},
  year={2024}
}

Exploring Semantic Variations in GAN Latent Spaces via Matrix Factorization ICLR 2023 Tiny Papers

Improved StyleGAN2 manipulations by replacing PCA with ICA..

Paper Code BibTeX

@misc{palaev2023exploring,
  title={Exploring Semantic Variations in {GAN} Latent Spaces via Matrix Factorization},
  author={Andrey Palaev and Rustam A. Lukmanov and Adil Khan},
  url={https://openreview.net/forum?id=2Z-dQTRezZ},
  year={2023}
}

Experience

Education

Tech Stack

Python PyTorch TensorFlow C++