Scholar
Mikolaj Czerkawski
Google Scholar ID: 6KVrra8AAAAJ
Partner Scientist, Asterisk Labs
Citations & Impact
Citations
212
H-index
9
i10-index
8
Publications
20
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Published papers such as 'COP-GEN-Beta: Unified Generative Modelling of COPernicus Imagery Thumbnails' (CVPR 2025 MORSE Workshop); 'MESA: Text-Driven Terrain Generation Using Latent Diffusion and Global Copernicus Data' (CVPR 2025 MORSE Workshop); 'Global and Dense Embeddings of Earth: Major TOM Floating in the Latent Space' (Pre-print); 'IceCloudNet: 3D reconstruction of cloud ice from Meteosat SEVIRI' (Artificial Intelligence for the Earth Systems); 'Major TOM: Expandable Datasets for Earth Observation' (IGARSS 2024; International Geoscience and Remote Sensing Symposium); 'Multi-Modal Convolutional Parameterisation Network for Guided Image Inverse Problems'; 'Exploring the Capability of Text-to-Image Diffusion Models with Structural Edge Guidance for Multi-Spectral Satellite Image Inpainting' (IEEE Geoscience and Remote Sensing Letters); 'Diffusion Models for Earth Observation Use-cases: from cloud removal to urban change detection' (Big Data from Space (BiDS) 2023); 'IceCloudNet: Cirrus and mixed-phase cloud prediction from SEVIRI input learned from sparse supervision' (Neurips 2023 Workshop Spotlight 'Tackling Climate Change with Machine Learning'); 'From LAION-5B to LAION-EO: Filtering Billions of Images Using Anchor Datasets for Satellite Image Extraction' (ICCV 2023 Workshop 'Towards the Next Generation of Computer Vision Datasets: DataComp Track'); 'SatelliteCloudGenerator: Controllable Cloud and Shadow Synthesis for Multi-Spectral Optical Satellite Images' (MDPI Remote Sensing). Published a course named 'DiffusionFastForward Course', containing a minimal and flexible implementation of diffusion models for image generation and translation.
Research Experience
- Co-founder and partner scientist at asterisk labs; Research fellow at the European Space Agency Φ-lab; Hosted and produced a technical episode for the satellite-image-deep-learning podcast; Guest on the satellite-image-deep-learning podcast discussing clouds and satellite imagery.
Background
- Research interests involve computer vision, signal processing, and Earth observation. Obsessed about open-source and large datasets. Expertise covers multi-modal learning, generative models, image synthesis and manipulation, image super-resolution, image-to-image translation, computer vision for remote sensing applications, and computer vision for radar signal processing.
Miscellany
- Participated in multiple technical podcasts, including discussing the need for open data for AI applications to EO and efforts in that direction with Alistair Francis on Robin Cole's YouTube podcast; appeared in the satellite-image-deep-learning podcast talking about clouds and satellite imagery and his SatelliteCloudGeneratorTool.
Co-authors