Export
Pablo Gomez
· 1 previous affiliation
- Industrial University of SantanderSince 2017
- MS student · Since 2022
- Undergrad student · 2017–2022
Research Overview
Pablo Gomez focuses on computational imaging optimization and the application of visible and near-infrared spectroscopy to agricultural soil analysis. His recent work addresses the end-to-end design of optical coding elements and develops automated protocols for acquiring spectral signatures. Currently, he investigates computational algorithms to improve the estimation of soil properties such as organic carbon content.
Research Trajectory
In 2024, Pablo Gomez's research output demonstrated a dual focus on computational imaging optimization and applied spectroscopy for agricultural monitoring. As a first author, he investigated the design of optical coding elements (OCEs) in computational imaging systems, where high-dimensional signals are sensed through post-processed coded projections. In Middle output regularized end-to-end optimization for computational imaging, Gomez addressed a critical flaw in standard end-to-end optimization pipelines. He identified that while the overall network may achieve acceptable training performance, the encoder layers modeling the OCE often perform poorly because traditional loss functions fail to account for intermediate representations. To resolve this, the research proposed a middle-output regularization approach to improve the physical design and reliability of optical coding elements.
Parallel to his work in computational imaging, Gomez has led research applying visible and near-infrared (VIS-NIR) spectroscopy to non-destructive soil analysis. Recognizing that spectral acquisition is highly susceptible to variations from human interaction, bias, and illumination inconsistencies, he focused on developing standardized methodologies for agricultural applications. In Design and Implementation of an Automated Protocol for Spectral Signatures Acquisition on Colombian Agricultural Soil Samples Into the Visible and Infrared Range, he introduced a cost-effective, automated protocol designed to reduce these variations. This work aimed to improve the reliability of spectral data fed into machine learning algorithms, specifically targeting the continuous monitoring and sustainable management of Colombian agricultural soils.
Current Directions
The most recent papers in the evidence indicate an ongoing focus on refining computational methods for environmental and agricultural spectroscopy. Recent work investigates targeted algorithmic approaches to extract specific, actionable soil metrics from spectral data. Building on his automated acquisition protocols, Gomez has explored feature selection techniques to isolate the most informative spectral bands. In Computational algorithm for soil organic carbon percentage estimation through NIR spectroscopy, he developed an algorithm utilizing spectral selection within the near-infrared range to accurately estimate the percentage of soil organic carbon. This approach demonstrated significant performance improvements over traditional estimation methods, highlighting a continuing interest in bridging advanced computational algorithms with practical agricultural sensing.
Profile updated
Loading new papers…
Loading research neighbors…
See something wrong? Report a mistake