Pan detection system in real time healthcare environment
Versandfertig in 1-2 Wochen
29,99 €
inkl. MwSt.
PAYBACK Punkte
15 °P sammeln!
The negative feeling of pain is often involuntarily expressed through facial expressions. Facial expression therefore is an important non-verbal cue to determine if a person is in pain. This property can be applied for diagnosis of pain especially among patients who are differently newlinechallenged and lack the ability of expressing their issue. In spite of the developments made so far, this field still lags behind in finding pain expressing faces in an uncontrolled environment through unprocessed newlinereal time images and videos. To bridge this gap, the study proposed a hybrid or fusion mo...
The negative feeling of pain is often involuntarily expressed through facial expressions. Facial expression therefore is an important non-verbal cue to determine if a person is in pain. This property can be applied for diagnosis of pain especially among patients who are differently newlinechallenged and lack the ability of expressing their issue. In spite of the developments made so far, this field still lags behind in finding pain expressing faces in an uncontrolled environment through unprocessed newlinereal time images and videos. To bridge this gap, the study proposed a hybrid or fusion model that could adequately detect a face expressing pain. The model was executed with inputs taken from pre-recorded or stored newlinevideos and live streamed videos. It involved the combination of Patch Based Model (PBM), Constrained Local Model (CLM), and Active newlineAppearance Model (AAM) in concurrence with image algebra. This allowed the efficient pain identification from raw home-made stored newlinevideos and live stream even through a bad recording device and under poor illumination. The hybrid model was implemented in a frame-by-frame manner for feature extraction and pain detection. The feature extraction part was done in pixel-based and point-based representation. For point-based representation, a concept called image algebra was used. For classification, three approaches viz. histogram technique, Feed newlineForward Neural Network (FFNN), and Multilayer Back Propagation Neural Network (MLBPNN) were implemented and analyzed. The videos newlineof different subjects showed facial expressions of pain::face, not::pain face and neutral::face. A home-made dataset was produced for storing the videos which was later used as the input and the selected features were stored. This dataset served as the training set for the proposed model. Though the data was not highly sensitive it was sufficient to confer adequate information for detecting pain expression.
Andere Kunden interessierten sich für
