Carlos Sarmiento

Papers

Total Citations

H-Index

About

Carlos Sarmiento’s research lies at the intersection of computer vision, probabilistic modeling, and autonomous robotics, with a focus on enabling machines to perceive and navigate 3D environments intelligently. His major contributions include pioneering the use of Hidden Markov Models (HMMs) for visual recognition tasks, where he demonstrated that treating 3D point cloud features as sequential observations can effectively solve place recognition problems. In his most cited work, he systematically compared discrete HMMs against Convolutional Neural Networks for image classification, showing that spatial sequencing of visual fragments can rival deep learning approaches in specific contexts. More recently, Sarmiento introduced Sparse-Map, an unsupervised learning technique that automatically generates topological maps from sensor data, addressing the memory and scalability limitations of traditional 2D occupancy grids for 3D navigation. While his citation counts (4, 3, and 2 for his top papers) reflect an emerging career, his work is notable for creatively adapting classical probabilistic models to modern perception challenges. Sarmiento’s research offers a compelling alternative to purely deep learning-based approaches, making him a promising voice in efficient, interpretable robotic perception.