Overview
Activity Profiling for Minimially Invasive Surgery
Minimially Invasive Surgery (MIS) or key hole surgery can greatly reduce the trauma of the patients and shorten the recovery period. However, the complexity of the instrument controls, restricted vision and mobility, difficult hand-eye coordination, and lack of tactile preception of the MIS operations require a high degree of dexterity of the operator. Existing research has shown that it is important to provide task-oriented analyses that capture the complexity of relationships between perceptual-motor, spatial, and experiential factors in addition to external factors such as experience and coordination of opreating team and the quality of equipment used. All these are important to the investigation and priorisation of the education needs of the surgeons at all levels with a view to establish preferred models for structured education and training.
The objective of this work is to investigate the use of computer vision for profiling the activities (in terms of instrument tissue operation, operating team dynamics, etc) during minimal invasive surgeries such that key operational manoeuvres can be tracked and analysed.
More details...
Body Sensor Networks
e-AR (ear-worn Activity Recognition) Sensor
The design of the e-AR sensor was inspired by the human inner ear. The human inner ear consists of an auditory system (the cochlea) and a balancing (vestibular) system. Within the vestibular system there are two sensory mechanisms, called the semicircular canals and the otoliths for sensing the rotational and translational motions. To emulate the sensory functions of the human vestibular system, the e-AR sensor is equipped with a MEMS (Micro Electro-Mechanical System) 3-axis accelerometer which is capable of detecting acceleration in 3 dimensions (up and down, left and right, back and forth). An accelerometer consists of a mass, and when the sensor is moved, the mass moves. Electronic sensing components determine the acceleration. By positioning the accelerometer on the ear, the e-AR sensor can pick up similar information to the vestibular system, and this records the posture and activities of the user.Pervasive Sensing for Sports TrainingIn situ measurements of athletes' physiological parameters during training and competitions are essential for identifying the underlying elements which affects sport performances. To enable real-time continuous measurements of athletes' performance indices during training and competitions, a number of pervasive sensing technologies have been introduced under the ESPRIT (Elite Sport Performance Research in Training) Programme. Through working closely with sports governing bodies, technologies developed have been validated for different sport exemplars. Under the programme, different novel sensing technologies have been introduced from body worn biomotion sensors, wheel chair velocity and tracking system, to rowing blackbox, and video tracking system.

Pervasive sensor for gait analysis
Gait analysis is an important part of orthopedics, rehabilitation, sport medicine and biomechanics. To quantify gait, motion capture systems, electromyography (EMG) sensors and force plates (or foot pressure insoles) are commonly used to capture kinematics, muscle contraction, and Ground Reaction Force (GRF). To measure GRF, force plates or foot pressure insoles are commonly used; however, both systems are costly, thus limiting their use mainly to dedicated biomechanics laboratories. The measurements performed are also constrained to brief time periods which may or may not represent the normal walking/running conditions. A novel concept of an ear-worn sensor is introduced for pervasive sensing of GRF patterns, and a hierarchical Bayesian network is developed to estimate the planar force distribution from the raw e-AR sensor signals. The approach has been validated against commercially available foot pressure sensing insoles and it has been shown that the sensor can accurately estimate the plantar force distribution.

More details...