Computer Vision Department

The Computer Vision Department teaches the full stack of core techniques - CNN-based image processing, segmentation with U-Net and SegFormer, object detection, and 3D point-cloud processing from LiDAR and photogrammetry. Students apply these methods directly to geoscience data through project work on satellite imagery, seismic visualization, and geological image interpretation, connecting theory to practice. By graduation, students can independently build vision pipelines that extract geologically meaningful information from raw imagery and point-cloud data.

⚙️ GeoAcademy System Administration College 🖥️ AI & Computing Sciences College 🔬 Basic Sciences College ⚡ Intelligent Geophysical Exploration College 🛢️ Resource & Energy Engineering College 🌍 Applied Geoscience Solutions College 💼 Economics, Policy & Strategy for the Future College 🎓 Education & Training Development College 🚀 Innovation & International Collaboration College

Machine Learning Department Computer Vision Department Natural Language Processing Department High-Performance Computing Department Quantum Computing Department Generative AI & Foundation Models Center Reinforcement Learning & Agents Center AI Safety & Alignment Center AI for Science Center Robotics & Embodied AI Center

🤖

🔑

⭐ LeCun

Chair

💡 Convolutional Neural Networks & Visual Recognition

Researchers (real GeoVerse Lab members) 15

🔑

1926–2013

David H. Hubel

Researcher

💡 Discovered (with Torsten Wiesel) the hierarchical organization of the visual cortex — simple, complex, and hypercomplex cells building increasingly abstract representations — the direct biological blueprint for convolutional neural network architecture

🧑‍🔬

🔑

1930–1985

King-Sun Fu

Researcher

💡 Founded syntactic pattern recognition, treating complex patterns as structured compositions of simpler primitives via formal grammars, and co-founded the field's flagship journal and society, earning the title 'father of pattern recognition'

🔑

1926–2008

Oliver Selfridge

Researcher

💡 Proposed the Pandemonium model — competing 'demons' detecting features that vote toward pattern recognition decisions — an early feature-voting architecture directly ancestral to modern object detection's feature-then-classify pipeline

🔑

1931–2004

Azriel Rosenfeld

Researcher

💡 Founded digital image processing and image segmentation as a rigorous discipline, writing the field's first textbook and establishing thresholding, region-growing, and digital topology methods still foundational to segmentation pipelines

🧑‍🔬

🔑

1885–1967

Erwin Kruppa

Researcher

💡 Proved Kruppa's theorem (1913) establishing that relative camera motion and 3D structure can be recovered from a minimal set of point correspondences between two views, the founding result of epipolar geometry and multi-view 3D reconstruction

🧑‍🔬

🔑

1904–1979

James J. Gibson

Researcher

💡 Originated the concept of optical flow — the pattern of apparent motion of surfaces in a visual scene caused by relative motion between observer and scene — founding the ecological approach to visual perception that underlies all optical flow algorithms

🔑

1788–1867

Jean-Victor Poncelet

Researcher

💡 Founded projective geometry, establishing the mathematics of perspective, vanishing points, and cross-ratio invariants that underlie all camera calibration and 3D-from-2D reconstruction in computer vision

🔑

1858–1927

Carl Pulfrich

Researcher

💡 Invented practical stereophotogrammetry — measuring 3D structure from paired stereo photographs — founding the discipline of photogrammetry that underlies LiDAR-independent 3D point cloud reconstruction from imagery

🧑‍🔬

🔑

1918–2000

Evelyn Pruitt

Researcher

💡 Coined the term 'remote sensing' and, as a US Office of Naval Research program director, built and funded the interdisciplinary scientific field connecting aerial/satellite imagery analysis to geography and earth observation

🔑

1834–1918

Ewald Hering

Researcher

💡 Founded the opponent-process theory of color vision (red-green, blue-yellow, black-white channels), the physiological model underlying color-space representations (e.g. Lab, YCbCr) used throughout computer vision and image processing

🔑

1921–2007

Ronald N. Bracewell

Researcher

💡 Developed Fourier-transform image reconstruction methods (including the Bracewell-Riddle algorithm for deconvolution) originally for radio-astronomy imaging, foundational to modern image deblurring, deconvolution, and super-resolution

🔑

1924–1998

Allan M. Cormack

Researcher

💡 Developed the mathematical theory of reconstructing a cross-sectional image from a set of projections (the Radon-transform-based basis of computed tomography), founding tomographic reconstruction used across medical, seismic, and scientific imaging

🔑

1924–2011

Willard Boyle

Researcher

💡 Co-invented (with George E. Smith) the charge-coupled device (CCD), the digital image sensor that made all digital photography, satellite imaging, and computer vision on captured imagery physically possible

🔑

1880–1943

Max Wertheimer

Researcher

💡 Founded Gestalt psychology, establishing the principles (proximity, similarity, closure, continuity) by which the visual system groups elements into coherent wholes — the perceptual-organization foundation of scene understanding and semantic segmentation

🔑

1928–2003

Béla Julesz

Researcher

💡 Invented random-dot stereograms proving stereoscopic depth perception requires no monocular cues, and founded texture perception theory (textons) explaining what visual textures are pre-attentively discriminable — foundational to computational texture analysis and stereo matching