Gabriel Martín

University of Cambridge

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Antonio Plaza

Universidad de Extremadura

José M. Bioucas‐Dias

Instituto de Telecomunicações

José M. P. Nascimento

Universidade Estadual de Campinas (UNICAMP)

Javier Plaza

Universidad de Extremadura

Mark S. Humayun

Government College University, Faisalabad

Sergio Bernabé

Universidad Complutense de Madrid

Amir H. Kashani

Johns Hopkins Medicine

Luis Ignacio Jiménez

Universidad de Valladolid

Sergio Sánchez

Instituto de Salud Carlos III

Maciel Zortea

IBM Research - Brazil

Cooperative Institutions

Universidad de Extremadura

University of Southern California

Instituto de Telecomunicações

University of Cambridge

Columbia University

University of Lisbon

University of Wisconsin–Madison

Southern California Eye Institute

Universidad Complutense de Madrid

University of Almería

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Fast spatial-spectral preprocessing for endmember extraction and spectral unmixing using graphic processing units

Luis Ignacio Jiménez Gabriel Martín Sergio Sánchez Javier Plaza Antonio Plaza

Linear spectral unmixing consists on the identification of spectrally pure constituents, called endmembers and their corresponding proportions or abundances using a linear model. Traditionally, most of the attention has been focussed on the exploitation of spectral information when identifying a set of endmembers and, only recently, some techniques try to take advantage of complementary information such as the one provided by the spatial correlation of the pixels in the image. Computational complexity represents a major problem in most of these spatial-spectral based techniques, as hyperspectral images provide very rich information in both the spatial and the spectral domain. In this paper we provide a computationally efficient implementation of a spatial-spectral processing (SSPP) algorithm which can be used prior to endmember identification and spectral unmixing. Specifically we present an implementation optimized for commodity graphics processing units (GPUs), which is evaluated using two different GPU architectures from NVidia: GeForce GTX580 and GeForce GT740. Our experimental validation reveals that significant speedups can be achieved when processing hyperspectral images of different sizes.

Endmember

Spectral signature

10.1109/igarss.2016.7730835

Cite

Citations (2)

Performance benefits of white formulation HDPE geomembranes

E3S Web of Conferences (2024)

Marc Amtsberg Borja Nanton Gabriel Martín Ignacio M. García

High Density Polyethylene (HDPE) Geomembranes have been used in containment applications for the past 50 years. Over this period, Polyethylene Geomembrane formulations have changed markedly with a design need to contain higher risk contaminants stored in more challenging environments. One of the major product developments has seen the emergence of “White Top Layer” Geomembranes. Atarfil would concur, and our most recent immersion and exposure data is compelling. It is not simply the thermal behavior that is beneficial, the additive types required to meet the challenges of stabilizing white layers, also serve significantly better protection when subjected to UV and Immersion Testing. This is translating to significantly longer predictions of Geomembrane performance life.

Geomembrane

White (mutation)

High-density polyethylene

10.1051/e3sconf/202456926002

Cite

Citations (0)

NONINVASIVE ASSESSMENT OF RETINAL VASCULAR OXYGEN CONTENT AMONG NORMAL AND DIABETIC HUMAN SUBJECTS

Retina (2014)

Amir H. Kashani Gilberto Raul Lopez Jaime Saloomeh Saati Gabriel Martín Rohit Varma

This pilot study was aimed to demonstrate the clinical feasibility of using hyperspectral computed tomographic spectroscopy to measure blood oxygen content in human retinal vessels.All procedures were performed under a University of Southern California Institutional Review Board-approved protocol and after obtaining informed consent. Fifty-seven subjects with and without diabetic retinopathy were dilated for standard fundus photography. Fundus photographs and retinal vascular oxygen measurements (oximetry) were made using a custom-made hyperspectral computed tomographic imaging spectrometer coupled to a standard fundus camera. Oximetry measurements were made along arteries (Aox) and veins (Vox) within vessel segments that were 1 to 2 disk diameters from the optic disk.For all control subjects (n = 45), mean Aox and Vox were 93 ± 7% and 65 ± 5% (P = 0.001), respectively. For all diabetic subjects (n = 12), mean Aox and Vox were 90 ± 7% and 68 ± 5% (P = 0.001), respectively. In subjects with proliferative diabetic retinopathy, Aox was significantly lower, and Vox was significantly higher than other groups (85 ± 4% and 71 ± 4%, respectively; P = 0.04, analysis of variance). There was a highly significant difference in the arteriovenous difference between subjects with proliferative diabetic retinopathy and those in the control group (14 vs. 26%, P = 0.003).Hyperspectral computed tomographic spectroscopy is a clinically feasible method for measurement and analysis of vascular oxygen content in retinal health and disease. This study uses the techniques relevant to oximetry; however, the breadth of spectral data available through this method may be applicable to study other anatomical and functional features of the retina in health and disease.

Fundus camera

Fundus (uterus)

Fundus Photography

10.1097/iae.0000000000000146

Cite

Citations (37)

Structural Health Monitoring on Urban Areas by Using Multi Temporal Insar and Deep Learning

Gabriel Martín Sivasakthy Selvakumaran Andrea Marinoni Zahra Sadeghi Campbell Middleton

The recent advancements in machine learning techniques have opened the door for automatic large scale monitoring of the surface of the earth. For instance, they could be used in order to evaluate and assess civil infrastructures at scale, which is costly due to the fact that typically the existing methods rely on in-situ evaluation. Over the last decade Deep Learning technologies have risen as the state of the art methods for many different machine learning problems due to the fact that they can learn complex features and model complex non-linear behaviours. In this paper we will explore the possibility of using Deep Learning technologies over remote sensing data with the aim of structure health monitoring at scale. We will compare the performance of new Deep Learning technologies with regards to other traditional machine learning methods. For this purpose, we will use InSAR (Interferometry Synthetic Aperture Radar) data which allow us to measure cumulative surface displacement in the line of sight of the sensor with millimetric accuracy. We will analyse multi temporal InSAR data in order to model ground subsidence. In this paper we will discuss how deep learning technologies can learn to detect terrain subsidence over multi-temporal InSAR data automatically, providing much better results than traditional methods.

Earth observation

10.1109/igarss47720.2021.9554639

Cite

Citations (3)

On Endmember Identification in Hyperspectral Images Without Pure Pixels: A Comparison of Algorithms

Journal of Mathematical Imaging and Vision (2011)

Javier Plaza E.M.T. Hendrix I. García Gabriel Martín Antonio Plaza

Endmember

Spectral signature

Simplex

Identification

Imaging spectrometer

10.1007/s10851-011-0276-0

Cite

Citations (90)

Determination of Spectral-Spatial Resolution of Hyperspectral Images For Retinal Imaging Applications

Justin S. Baba Amir H. Kashani Thomas P. Karnowski Gabriel Martín Mark S. Humayun

Full spectral imaging

Cite

Citations (0)

Hyperspectral Coded Aperture (HYCA): A New Technique for Hyperspectral Compressive Sensing

Zenodo (CERN European Organization for Nuclear Research) (2013)

Gabriel Martín José M. Bioucas‐Dias Antonio Plaza

Hyperspectral imaging is an active research area in remote sensing. Due to the high volume of hyperspectral image data, the exploration of compression strategies has received a lot of attention in recent years. In this paper, we introduce a new compressed sensing methodology, termed Hyperspectral coded aperture (HYCA), which exploits the high correlation existing among the components of remotely sensed hyperspectral data sets to reduce the number of measurements necessary to correctly reconstruct the original data. HYCA relies on two central properties of most hyperspectral images: i) the spectral vectors live on a low dimensional subspace and ii) the spectral bands are piecewise smooth. The former property allows to represent the data vectors using a small number of coordinates, and the latter implies that each coordinate is piecewise smooth and thus compressible on local differences. The reconstruction of the data cube is obtained by minimizing a convex objective function containing a data term associated to the compressed measurements and a total variation spatial regularizer. A series of experiments with simulated and real data show the effectiveness of the newly developed HYCA, indicating that the proposed scheme has a high potential in real-world applications.

Coded aperture

10.5281/zenodo.43598

Cite

Citations (8)

Hyperspectral coded aperture (HYCA): A new technique for hyperspectral compressive sensing

Gabriel Martín José M. Bioucas‐Dias Antonio Plaza

In this paper, we develop a new lossy compression framework for hyperspectral images, termed hyperspectral coded aperture (HYCA), which combines the ideas of spectral unmixing and compressive sensing. It takes advantage of two main properties of hyperspectral data, namely the high spatial correlation that can be observed in the data and the generally low number of endmembers needed in order to explain the data. In other words, our proposed approach intends to exploit the fact that the high dimensional hyperspectral data lives in a subspace of much lower dimension due to the mixing phenomenon. Our experimental results, conducted with synthetic hyperspectral data, indicate that the proposed approach represents a promising new strategy.

Lossy compression

Full spectral imaging

Chemical Imaging

Aperture (computer memory)

10.1109/igarss.2012.6351279

Cite

Citations (18)

Retinal Vascular Oxygen Saturation in Diabetic Rats: A Study using Hyperspectral Computed Tomographic Imaging Spectroscopy

Investigative Ophthalmology & Visual Science (2015)

Ray Zhang Qianhaozhe You Gabriel Martín Mark S. Humayun Amir H. Kashani

Purpose: To assess changes in retinal vascular oxygen concentration in a diabetic rat model using hyperspectral computed tomographic imaging spectroscopy (HCTIS). Methods: Nine brown Norway rats were anesthetized and dilated for retinal imaging. Six rats were injected with 65 mg/kg of Streptozotocin (STZ) to induce hyperglycemia, and 3 rats with 0.9% saline control. Blood glucose levels were measured after a 12 hour fast using Accu-Chek Aviva Plus glucometer (Roche Inc). Baseline oximetry measurements and fundus photographs were taken in both eyes of each rat. HCTIS oximetry measurements were taken every 2 weeks after STZ injection for up to 2 months. The same vessel areas were selected for analysis at each time point. Fundus photos and FA were repeated at 2 months. Comparison of retinal arterial (A_(ox)) and venous (V_(ox)) oxygen saturation levels were made at each time point using the Student T-test. Results: All experimental rats developed and maintained hyperglycemia (mean 410±41 mg/dl), and all control rats remained normoglycemic (mean 124±24 mg/dl) over the 2 months of the study. For control rats (n=3) mean A_(ox) was 95±1%, 100±4%, 97±6%, 102±5%, and 100±7% and mean V_(ox) was 90±4%, 88±3%, 86±3%, 87±2%, and 87±5% at baseline, 2, 4, 6, and 8 weeks. For experimental rats (n=6) mean A_(ox) was 91±2%, 94±4%, 91±4%, 94±3%, 94±5% and mean V_(ox) was 89±3%, 90±4%, 88±2%, 86±3%, 85±3% at baseline, 2, 4, 6, and 8 weeks. There were no significant differences between control and experimental A_(ox) and V_(ox) for each time point. The arteriovenous difference at each time point also showed no clear trends. V_(ox) measurements in the diabetic rats showed a decreasing V_(ox) trend over time when compared with baseline (p=0.76, 0.57, 0.17, 0.07). Conclusions: There was a trend towards decreasing V_(ox) over 2 months time in experimentally induced diabetic rats but this was not statistically significant. There were no significant differences in A_(ox) and V_(ox) in rats with experimentally induced diabetes versus controls. HCTIS is a useful method for measuring retinal vascular oxygen content in the rat model.

Fundus camera

Fundus (uterus)

Oxygen Saturation

Source

Cite

Citations (0)

Snapshot Hyperspectral Imaging: Retinal Oxygen Saturation Analysis From Hyperspectral Images of Rabbit Retina

Amir H. Kashani Gabriel Martín Amani A. Fawzi Daniel W. Wilson William R. Johnson

Snapshot (computer storage)

Cite

Citations (0)