Narrow Results

The quality of respiratory function determines the recovery and survival rate of the patients with cervical spinal cord injury. A cost efficient method of evaluating the respiratory function is to assess the strength of their cough sounds. However, some patients with cervical spinal cord injury fail to develop an effective cough because of pain or nerve damage, and their voice is shout, called pseudo-cough herein, rather than cough. Such samples of pseudo-cough sounds should be weeded out so as to avoid wrong evaluation of respiratory function. In this paper, a linear classifier is proposed to recognize pseudo-cough sounds from cough sounds for patients with cervical spinal cord injury. To alleviate dependence on the number of cough-sound and pseudo-cough-sound samples, a light-weight classifier is constructed by using merely two features: zero-crossing rate and maximal autocorrelation coefficient, and the classifier is trained mainly with unvoiced and voiced sounds rather than cough sounds and pseudo-cough sounds. Experimental results showed that the sensitivity and specificity were 98% and 86.4%, respectively.

Two significant problems in content based retrieval methods are (1) Accuracy: most of the current content based image retrieval methods have not been quantitatively compared nor benchmarked with respect to accuracy and (2) Efficiency: image database search methods must be analyzed for their computational efficiency and interrelationships. We assert that the accuracy problem is due to the generality of the applications involved. In the current systems, the goal of the user is not clear, which results in difficulties in creating ground truth. In this paper, we quantitatively compare and evaluate four fundamentally different methods for image copy location, namely, optimal keys, texture, projection, and template methods in a large portrait database. We discuss some important theoretical interrelationships, computational efficiency, and accuracy with respect to real noise experiments.