Maryam Haghighat

Commonwealth Scientific and Industrial Research Organisation, Queensland University of Technology

Papers

Total Citations

H-Index

About

Maryam Haghighat is a rising researcher at the intersection of computer vision, robotics, and remote sensing. Her primary research areas include hyperspectral image analysis and multi-robot localization in unstructured environments. Haghighat’s most notable contribution is the development of **FactoFormer**, a factorized hyperspectral transformer architecture that leverages self-supervised pretraining to efficiently capture both spectral and spatial dependencies in hyperspectral images. This work, published in 2023, has already garnered 30 citations, signaling its impact on advancing transformer-based methods for remote sensing. In parallel, Haghighat has tackled the challenge of robust multi-robot re-localization in natural environments, such as forests, where single-modality approaches often fail. Her 2023 paper on this topic, with 4 citations, proposes a deep learning framework that fuses multiple sensor modalities to achieve reliable place recognition and localization under adverse conditions. This work has practical implications for autonomous systems operating in GPS-denied or visually degraded settings. Haghighat’s ability to bridge theoretical innovation with real-world deployment—from satellite imagery to field robotics—marks her as a promising talent in embodied AI and geospatial intelligence.