My current research interests are in image processing, with particular focus on texture and colour analysis. In the past I worked on geometric modeling, CAD/CAE and simulation of complex systems. At present the research areas in which I'm active are:

Mathematics of textures

Texture analysis has been an area of intense research activity for more than three decades. In this context we are trying to define new texture descriptors based on mathematical concepts such are space partitioning an polytopes. We have recently showed that some well known methods (e.g.: Local Binary Patterns) are based on this idea.

The figure above reports three polytope graphs corresponding to the following local binary patterns (LBP3x3): 00000000, 00000101 and 00001101

In cooperation with:
  • Universidade de Vigo (Spain)
Related papers: JEI-2017, IIMSS-2017a, IS2016, CTMR2015b, PRL-2014b, JMIV-2014, SCI-2014, JMIV-2013, JMIV-2011, OLEN-2011, PR-2007

Colour texture analysis

Colour has proved to improve texture analysis in many contexts. We are concerned with the development of new image descriptors through combination of colour and texture. We proposed two new descriptors: colour ranklets and multilayer coordinated cluster representation.

In cooperation with:
  • Universidade de Vigo (Spain)
  • University of East Anglia (UK)
Related papers: JEI-2016, CTMR2015a, AOT-2013, ESWA-2013, ESWA-2012, JEI-2011, NCC-2011a

Medical image analysis

Discrimination between epithelium and stroma in hystological images
The tumour-stroma ratio (TSR) is an independent prognostic factor in a wide number of oncologic disorders. As a consequence, accurate and reproducible estimation of TSR plays a central role in patient stratification and follow-up. Manual assessment of TSR can be time consuming and may suffer from a set of drawbacks such as poor reproducibility and high inter- and intra-observer variability. The objective of this research is to develop automated or semi-automated approaches for computer-assisted estimation of the tumour-stroma ratio.

Epithelium Epithelium.jpg
Stroma images/Stroma.jpg

In cooperation with:
  • Universidade de Vigo (Spain)
  • Leiden University Medical Center (The Netherlands)
Related papers: IIMSS-2017b, SR-2016, NEUCOM-2015

Evaluation of shape and textural features from CT/PET/MRI as potential bio-markers for diagnosis, stratification and follow-up
We investigate textural features as potential imaging biomarkers, based on the assumption that they are an index of the degree of tumour heterogeneity.

CT scans

In cooperation with:
  • Perugia Hospital 'S. Maria della Misericordia' (Italy)
Related papers: AR-2018

Surface inspection & grading

Paper inspection
Paper may contain particles of various type, and the papermaking industry is increasingly concerned with the development of quick and reliable systems to detect and characterise such inclusions automatically. We studied a sequential, two-step procedure based on preliminary classification to differentiate defective paper patches from the defective ones followed by thresholding to separate the impurities from the background.

PaperInspection_2 PaperInspection_3 PaperInspection_4
Sample to analyse Inspection areas Detection of defective areas Separation of the impurities from the background

In cooperation with:
  • Universidade de Vigo (Spain)
Related papers: COMIND-2014, JSCOM-2014, EMSS-2013

Automated grading of natural stones

This research activity is about grouping product into lots based on the criterion of "similar appearance". In particular we are concerned with grading of natural stone products. We are building a database of granite tiles acquired under controlled illumination conditions (current version available here) and testing the effectiveness and robustness of various texture and colour descriptors.


In cooperation with:
  • Universidade de Vigo (Spain)
  • Mondial Marmi Srl (ITALY)
Related papers: ESWA-2012, MVA-2011, DYNA-2010

Remote sensing

Detection of coral reefs from sidescan sonar imagery
Sidescan sonar is commonly used to capture high-resolution acoustic imagery of the seabed. Different seabed regimes and habitats produce textural signatures recognisable to a human interpreter. Discriminating a particular habitat from the background can be thought of as a spatially distributed target recognition problem. We are currently investigating the effectiveness of texture analysis methods in this context.

In cooperation with:
  • University of East Anglia (UK)
Related papers: IMVIP-2011

Grain-size assessment of aggregates through area morphology
Grain-size analysis of unconsolidated particles plays an important role in many areas of science and engineering including sedimentology, waste management, nanomaterials, corrosion and food engineering. The objective of this research is the development of image-based methods for computing grey-scale granulometries and estimating the mean size of fine and coarse aggregates. We are currently investugating the use of area morphology and combined information from both openings and closings to determine the size distribution.

VeryCoarseSand.jpg VeryFinePebbles.jpg FinePebbles.jpg MediumPebbles.jpg CoarsePebbles.jpg

In cooperation with:
  • Universidade de Vigo (Spain)
  • University of East Anglia (UK)
Related papers: MVAP-2015, WM-2016

Greenhouse detection and classification from satellite images
Plastic-covered agriculture is part of the transformation of the conventional farming into a more industrial and 'high-tech' agriculture. Unfortunately, it is well knwon that plasticulture has significant anthropic impact: careful spatial development planning is therefore required to minimise the potential downsides of this type of cultivation. Satellite imagery represents a powerful means for local governments and environmental agencies to monitor the use of soil and greenhouse development. Our investigation involves the use of image stereo-pairs from GeoEye-1 and WorldView-2 for object-based detection and classification of plastic- or net- covered greenhouses.


In cooperation with:
  • Universidad de AlmerÝa (Spain)
Related papers: RS-2014

Other applications

Analysis of urea-water sprays through image processing
The injection of urea-water sprays within selective catalytic reduction systems is the leading technique for reducing the emission of nitrogen oxides from Diesel engines. For the process to work properly, it is crucial to guarantee the adequate size, velocity and distribution of the spray droplets upstream of the catalyst. It is therefore extremely important to understand the process of spray formation and evolution as well as possible. The objective of this reaserch is to develop a methodology for inspecting the behaviour of urea-water sprays in realistic conditions. Our interest is in invetistigating the use of imaging-based method as potential alternatives to standard methods such as phase-Doppler anemometry.

Urea-water spray (Original image, backlight)
Image after background removal
Binarized image

In cooperation with:
  • Spray Lab (UniversitÓ degli Studi di Perugia, Italy)
Related papers: FUEL-2015

Biometric identification through skin texture features
It is well known that skin texture is a relevant source of information for biometric identification. Specifically, in this activity we investigated the role of skin texture features from the forearm. This has a number of potential advantages in that the forearm is a relatively easy-to-access area and normally shielded from the elements and wear to which other parts of the body are exposed. We proposed a method based on double-imaging modality (optical + capacitive) which requires neither registration of the optical and capacitive images nor that they represent the same patch of skin between them or across different acquisitions.

Skin-1 Skin-2 Skin-3

In cooperation with:
  • Queen Mary, University of London (United Kingdom)
  • London South Bank University (United Kingdom)
Related papers: SRANDT-2017, IPTA-2017