In this research, an algorithm to automatically detect artery wall region and estimate systolic and diastolic lumen diameter of the artery is proposed using ultrasound radio-frequency signals acquired with a single ultrasound transducer in M-mode.Detection of the region containing the artery walls is obtained using a machine learning trained classifier. Artery lumen diameter is estimated using peak detection and correlation technique.The proposed algorithm is first tested on M-mode signal acquired with human in-vivo experiments using a clinical ultrasound imaging system and a wearable ultrasound sensor (WUS).The machine learning-based artery wall region classifier was able to detect the artery wall region in M-mode signals acquired from the clinical ultrasound acquisition system with an accuracy of 88.9%. In the case of data acquired from WUS, it is 88.3%.Further, the diameter estimated with the proposed technique was within the range of the average diameter of the common carotid artery observed in humans.