Purpose: Anterior Cruciate Ligament (ACL) injury is among the most prevalent knee injuries among young athletes, aged 15-25 and occurs in 70% through a noncontact mechanisms. Medial Knee Displacement (MKD) is the key risk factor of ACL injury in basketball athletes. The present study investigated the effect of an MKD corrective exercises program on knee kinematics during Drop Vertical Jump (DVJ...

Even though many approaches have been proposed for automated fault detection and diagnosis of Heating Ventilation and Air Conditioning (HVAC) systems, and advanced maintenance programs, such as preventive and predictive maintenance have been applied for the systems, issuing work orders for corrective maintenance is still a reality. Currently corrective maintenance still accounts for more than 5...

BACKGROUND Training bodies see teaching by junior doctors and vocational trainees in general practice (family medicine) as integral to a doctor's role. While there is a body of literature on teacher training programs, and on peer and near-peer teaching in hospitals and universities, there has been little examination of near-peer teaching in general practice. Near-peer teaching is teaching to th...

Despite the importance of behaviour management training, many student teachers report being inadequately trained in this area. The aim of this study was to identify the strategies, confidence and reported levels of success in regards to various behaviour management strategies, across 509 first, second, third and fourth year student teachers training to be primary teachers. The most significant ...

In this work the output density functions of hidden Markov models are phoneme-wise tied mixture Gaussians. For training these tied mixture density HMMs, modiied versions of the Viterbi training and LVQ based corrective tuning are described. The initialization of the mean vectors of the mixture Gaussians is performed by rst composing small Self-Organizing Maps representing each phoneme and then ...

This work presents training methods and recogni tion experiments for phoneme wise tied mixture den sities in hidden Markov models HMM The system trains speaker dependent but vocabulary independent phoneme models for the recognition of Finnish words The Learning Vector Quantization LVQ methods are applied to increase the discrimination between the phoneme models A segmental LVQ training is pro p...

CAVE-like displays allow a user to walk in to a virtual environment, and use natural movement to change the viewpoint of virtual objects which they can manipulate with a hand held device. This maps well to many surgical procedures offering strong potential for training and planning. These devices may be networked together allowing geographically remote users to share the interactive experience....

BACKGROUND Posture instability and unsteady gait disorders in Parkinson's Disease (PD) usually contribute to fall-related fractures. Fall-related trauma in PD is the most common reason for injury. Despite providing modern care for PD patients (PP) in the recent years, anti-PD drugs have no effect on falling. There is an urgent need to administer exercise interventions to reduce falls and relate...

SRI has developed the DECIPHER system, a hidden Markov model (HMM) based continuous speech recognition system typically used in a speaker-independent manner. Initially we review the DECIPHER system, then we show that DECIPHER's speakerindependent performance improved by 20% when the standard 3990-sentence speaker-independent test set was augmented with training data from the 7200-sentence resou...