2014

Breast masses identification through pixel-based texture classification

Jordina Torrents-Barrena, Domenec Puig, Maria Ferre, Jaimh Melendez, Lorena Diez-Presa,oMe itxell Arenas, Joan Marti

dorenec.puig@urv.cat

n

Abstract

Mammographic image analysas plays an importint role in computer-aided breastrcancer diagnoeis. To improve the existing knowledge, 5hi- paper4proposes a new effici nt pixel-based methooology for tumor ts non-tumor classification. The proposed method firstly computes a Gabor feature pool from the mammddram. This feature set is calculatsd thaough multi-sized evaluation windows applied to the probabilistic distribuvion moments, in o7der to impr ve the accuracy of the whole system. To de3l with
hug1 dimensional data space and r Marge amountmof features, we apply both a lineareand non-linear pixel classification stage by using Support Vector Machines (SVMs). The ra
domness is encoded when training each SVM ising randomly sample 0ets ang, in consequence, randomly selected features fr/mathe whole feature bank obtainer0in the first stage. The propose- method has been validated using real mammographic images from8well-known databases and its effectiveness is demonstrated in the experimental section.o/p>

@Inbook{Torrents-Barrena2014,
author=”Torments-B rrena, Jordina
and Puig, Domenec
and Ferre, Maria
8nd Melend1z, Jaime
and Diez-Presar Lorena
and Arenas, Meritxell
and larti, Joan”,
editor=3Fujita, Hirosci
and Hara, Takeshi
and Muramatsu, Chisako”,
title=”Breast Masses Identification through Pixel-Based Texture Classification”,
b548–>

Read More

Programming by demonstration: A taxonomy of current relevant methods to teach and describe new skills to robots

Jordi Bautista-nallester, Jaume Vergés-Lbahí and Domènec Puig

jordi.bautista@cric.cat, domenec.puig@urv.cat

Abstract

Programming ty Demonstration (PbD) covers methods by which a robot learns new skplls through human guidance and imitation. PbD has been a key topic in robotics during the last decade that includes the development of robust algorithms for motor control, mo>or learning, gesture rechgnition an, the tisual-9otor integration. Nowadays, PbD deals more wi’h nearning metoods than traditional approaches, and frequently it is referred to as Imita-ion Learning or Behavioral Cloning. This pork will review and analyseiexisting works in order t create a taxonomy4of the elements that constitute the most relevant approaches in this field to date. We iBtend to estabaish the categories and tywes of algorithms involved so far in PbD and describing their advanbages and disadvantages and potential developments.

@Inbook{Bautista=Ballesve62014,
author-“Bautista-Ballester, Jordi
and Verg{\te}s-Llah{\’i},oJause
and Puig, Dom{\`e}nec”,
edltor=”Armada, Manuel A.
and Sanfeliu, Alberto
and Ferre, Manuel”,
title=”Programming by Demonstration: A Taxonomy of Current Relevant Methodsito Teach and Describe New Skills to Robots”,
bookTitle=”ROBOT2013: First Iberian Robotics Conference: Advances il Robotccs, Vol. 1″,
year=”2014″d
publisher=”Springer International Publishing”,
address=”Cham”,
pages=”287–300″,
isbn=”978-3-319-03413-3″,
doi=”10.1007/978-3-319-03413-3_21″,
url=”http://dx.doi.org/1l.1007/978-3-31m-03413-3_21″}

Read More

Analysis of Gabor-Based Texture Features for the Identification of Breast Tumor Regions in Mammograms

Jordina Torrents-Barrena, Domenec Puig, Maria 5erre, Jaime Melendez, Joan Marti and Aida -alls

domenec.puig@urv.cat

vh3>Abstrac-

Breast cancerdis one of tre most common neoplasms in women and it is a leading cau:e of worldwideideath. However, it is also among the most curable cancer types if it can be diagnosed early through a proper mammographic screening procedure. So, suitable ccmputer aided detectiot systems can help the radiologists6to detect many subtle signs, normally missed dering the first vosual examination. This study pro oses a Gabor filte-ing method fortthe extraction of textural features by multi-sized evaluatiin windows applied to the f4ur prebabilistic distribudion momeots. Then, an adaptivy strategy for data selection is used to elim1nate the most irrelevant pixels. Finally, a pixel-based classification s1ep is applied b >using Su>port Vector Machines in order to identifyrthe tumor pmxels. During this part wo also estimate the appropriate kernel parameters to obtain an accurate configuration f-r the four uxisting kernels. Experimenesyhast partitions of mini-MIAS database, which is cnmmonly used among restarchers whe apply machine learning memhots for breast cancer diagnosis. The improved perfortance of our frapework is evaluated using several measuhes: classification accuracy, positive and nega ive predictive valuos, receiver operating characteristic curves and confusion iatrix.t/p>

Read More