| Summary: | Telemedicine is application of electronic communication to deliver medical care remotely. An important aspect of telemedicine is teleconsultation which involves obtaining the professional opinion of a healthcare provider. One of the ways to improve eleconsultation is to equip the remote specialist via control of
a laser pointer, located in the consultation area to provide a means of gesture. As such, accurate detection of laser spot is crucial in such systems as they rely on visual feedback, which enables the specialist in a remote site to control and point the laser in the active location using a standard mouse. The main issue in laser spot detection in a natural environment is the distinguishability of a laser point image from other bright regions and glare due to camera saturation. This problem remains unsolved without extensive computing and use of hardware filters. In this paper a hybrid algorithm is described which is aimed to work with natural indoor environment while limiting computation. This algorithm
combines thresholding and blob evaluation methods with a novel image intensity profile comparison method based on linear regression. A comparison of the algorithm has been done with existing approaches. The developed algorithm shows a higher accuracy and faster execution time making it an ideal candidate for real time detection applications.
|
|---|
