Heart rate measurement and blur image quality metric via image processing: An overview

The research area of image processing and its applications has significantly increased in the last decades. They can be found in medical imaging, military, surveillance, astronomy, sports, and many more. In this paper, we show two applications of image processing; heart rate measurement and a no-reference blur image quality metric. The heart rate measurement is obtained from video sequences. However, the heart rate measurement consists of artifacts and noises. Therefore, a novel method using the Canonical Component Analysis (CCA) is introduced to eliminate the artifacts and noises using two five-second video frames where one video is delayed by one second. In a no-reference blur image quality metric, we show that some exact Zernike moments (EZMs) closely exhibit the human visual system in assessing the quality of images distorted by various degrees of Gaussian blur. These EZMs are then combined through a formulation and trained using a support vector machine to give the quality scores for images from the widely used public databases.