Tsai Y.-W.JIAN-JIUN DING2022-04-252022-04-252021https://www.scopus.com/inward/record.uri?eid=2-s2.0-85123490349&doi=10.1109%2fGCCE53005.2021.9622011&partnerID=40&md5=c44ec0bc61b745eafb1685ead10f3eechttps://scholars.lib.ntu.edu.tw/handle/123456789/607204Image sharpness is a metric to determine whether a region within an image is clear or blurring. It is helpful for image quality measurement, foreground extraction, image enhancement, depth estimation, and adaptive image processing. However, for most of existing sharpness measurement methods, it costs much computation time to compute the sharpness locally and pixelwisely. In this paper, we propose an effective and efficient image sharpness measurement method. Its idea is based on that, for a blurred edge, the response of the short-length edge detector is much smaller than that of the long-length edge detector. By contrast, for a sharp edge, the difference is relative smaller. Experiments show that, with the proposed method, the blurred and the clear regions can be well distinguished. Moreover, it requires much less computation loading and can be performed in real time. ? 2021 IEEE.blurringedge detectorimage metricsharpnessImage qualityBlurringDepth EstimationEdge detectorsForeground extractionImage metricsImage quality measurementImage sharpnessLongest edgeMeasurement methodsSharpnessImage enhancementconference paper10.1109/GCCE53005.2021.96220112-s2.0-85123490349