Automatic Analysis of Digital Retinal Images for Glaucoma Detection

In this paper we propose a novel automated glaucoma detection framework for mass-screening that operates on inexpensive retinal cameras. The proposed methodology is based on the assumption that discriminative features for glaucoma diagnosis can be extracted from the optical nerve head structures,
such as the cup-to-disc ratio or the neuro-retinal rim variation. After automatically segmenting the cup and optical disc, these features are feed into a machine learning classifier. Experiments were performed using two different datasets and from the obtained results the proposed technique provides
better performance than approaches based on appearance. A main advantage of our approach is that it only requires a few training samples to provide high accuracy over several different glaucoma stages.

An anatomical study of the arterial and nerve supply of the infrahyoid muscles

A precise knowledge of the sources of the arterial and neural supply of the sternohyoid (SH), sternothyroid (STM), and superior belly of omohyoid (OM) is of value to surgeons using the infrahyoid muscles in reconstruction procedures of the head and neck. This study was designed to define the anatomical bases of the variable sources of the arterial and neural supply of these muscles. Fourteen cadavers were unilaterally dissected in the neck region, and the arterial pedicles of these muscles were followed and accurate measurements were taken. For the SH, two arterial pedicles (superior and inferior) originated from the superior thyroid artery ST and supplied the muscle in 57.1% of cases. The inferior pedicle was absent in 42.9% of cases. As regards the STM, one arterial pedicle from the ST supplied its upper end by multiple branches in 57.1% of cases. In 14.3% of cases, branches from the inferior thyroid artery (IT) supplied the STM in addition to its supply from the ST. As regards the OM, two arterial pedicles originated from the ST and supplied its upper and lower ends in 57.1% of cases. The main artery from the ST to the superior belly of OM entered at its superior portion. The ansa cervicalis (AC) innervated the infrahyoid muscles. SH usually had a double nerve supply. In 57.1% of cases, its superior part was innervated by the nerve to the superior belly of OM. Its inferior part received branches from the AC. In 35.7% of cases, its superior part received direct branches from the AC. As regards the STM, in (71.4%) of cases, a common trunk arose from the loop and supplied the inferior part of both the SH and STM. The nerve supply to the superior belly of OM originated from the AC below the loop in 64.3% of cases. These data will be useful for preserving the neuro-vascular supply of the infrahyoid muscles during flap preparation.

To be inside someone else's dream: On Music for Sleeping & Waking Minds

Music for Sleeping & Waking Minds (2011-2012) is a new,overnight work in which four performers fall asleep while wearing custom designed EEG sensors which monitor their brainwave activity. The data gathered from the EEG sensors is applied in real time to different audio and image signal processing functions, resulting in continuously evolving multichannel sound environment and visual projection. This material serves as an audiovisual description of the individual and collective neuro physiological state of the ensemble. Audiences are invited to experience the work in different states of attention: while alert and asleep, resting and awakening.

Forward and backward least angle regression for nonlinear system identification

A forward and backward least angle regression (LAR) algorithm is proposed to construct the nonlinear autoregressive model with exogenous inputs (NARX) that is widely used to describe a large class of nonlinear dynamic systems. The main objective of this paper is to improve model sparsity and generalization performance of the original forward LAR algorithm. This is achieved by introducing a replacement scheme using an additional backward LAR stage. The backward stage replaces insignificant model terms selected by forward LAR with more significant ones, leading to an improved model in terms of the model compactness and performance. A numerical example to construct four types of NARX models, namely polynomials, radial basis function (RBF) networks, neuro fuzzy and wavelet networks, is presented to illustrate the effectiveness of the proposed technique in comparison with some popular methods.