Maria Luisa Durán

Recognizing marbling in dry-cured Iberian ham by multiscale analysis

Dry-cured Iberian ham is one of the most valuable meat products in Spain, with a first-rate consumer acceptance. Visually discernible characteristics of fat and lean, such as marbling, have all effect on the acceptability and palatability of ripened Iberian ham pieces. Important marbling properties include the amount and spatial distribution of intramuscular fat streaks. Chemical processing is the only proved way to determine the fat level of pig meat, but this technique is tedious, destroying and unable to offer information about fat distribution. The determination of Iberian ham sensorial quality has traditionally involved appraisal of marbling characteristics by descriptive analysis methods, which rely heavily on visual evaluation and testing by panels of trained graders. We present a novel method to recognize marbling in Iberian ham images to provide the base for the design of an automatic, non-destroying expert computer system, based on computer vision and pattern recognition techniques, which shall allow food technology industries to evaluate and characterize Iberian ham independently of the subjective and variable criteria of human testers.

Optimization of MRI Acquisition and Texture Analysis to Predict Physico-chemical Parameters of Loins by Data Mining

The main objective of this study was to configure the acquisition and analysis of low-field magnetic resonance imaging (MRI) to predict physico-chemical characteristics of Iberian loin, evaluating the use of different MRI sequences (spin echo, SE; gradient echo, GE; turbo 3D, T3D), computational texture feature methods (GLCM, NGLDM, GLRLM, GLCM + NGLDM + GLRLM), and data mining techniques (multiple linear regression, MLR; isotonic regression, IR). Moderate to very good correlation coefficients and low mean absolute error were found when applying MLR or IR on any method of computational texture features from MRI acquired with SE or GE. For T3D sequence, accurate results are only obtained by applying IR on GLCM or GLCM + NGLDM + GLRLM methods. Considering not only the accuracy of the methodology but also consumed time and required resources, the use of SE sequences for MRI acquisition, GLCM method for MRI texture analysis, and MLR could be indicated for prediction physico-chemical characteristics of loin.

Behaviour of Texture Features in a CBIR System

Searching and processing in databases of general and non-specific images are highly subjective. The process of texture feature extraction from images produces results of highly theoretical and mathematical character that have little to do with human perception. We present a method to select from low-level texture features, statistics and numerical groupings and to transform them into other high-level features, with visual meaning. We also aim to facilitate their use within CBIR systems. The detailed study of the composition and behaviour of the texture characteristics has enabled us to abstract and use them in an automated manner, similarly to how an observer would do.

3D Reconstruction on MRI to Analyse Marbling and Fat Level in Iberian Loin

Dry-cured Iberian pig products are some of the most valuables meat products in Spain. Visually discernible features of fat and lean, such as marbling, have an effect on the acceptability of these products. Marbling properties include the amount and spatial distribution of intramuscular fat streaks. Thresholding techniques are the simplest and most widely used for image segmentation, but pay no attention to all spatial information on the images. In this paper we propose another method to evaluate fat level in loin and make a three-dimensional visualization with only the isolated fat in order to give knowledge about fat distribution, and about width of fat streaks, helping the experts in food science to analyze marbling. In order for fat isolation we apply a method based on a pyramidal decomposition of images combined with region-growing techniques. 3D reconstruction is obtained by marching cubes algorithm.

Testing Geodesic Active Contours

Active Contours are a widely used Pattern Recognition technique. Classical Active Contours are curves evolutionate by minimizing an energy function. However, they can detect only one o bject within an image with several objects, and the solution is highly dependent on parameters in its formulation. A solution can be found in Geodesic Active Contours (GAC). We have developed a version of this technique and improved some aspects to apply on real and practical cases. The algorithm has been tested with both synthetic and real images.

An Automatic Non-Deforming Technique for Selective Microcalcifications Enhancement

Breast cancer is the leading cause of death among middle aged and older women. Microcalcifications are one of the earliest cancer signs, which are an important sign of up to half of occult breast carcinoma.