Heart failure (HF) has been identified as an epidemic affecting over 23 million worldwide, due to its associated significant mortality, morbidity, and healthcare expenditures [1]. Acute heart failure is the leading Medicare discharge diagnosis that leads to readmission within 30 days and a 5-year mortality rate of approximately 42% [2]. Hospital admission for HF is one of the indicators of prognosis, as one prospective study of 257 patients admitted to the hospital non-electively for HF who were discharged alive. Of those 257 patients, 82 (32%) died or were readmitted to the hospital within 60 days of discharge [3]. Traditional physical exam findings are insensitive to pulmonary edema and newer hemodynamic monitors are invasive, expensive and not widely available for patients with HF. The use of lung ultrasonography has been shown to be effective in detecting pulmonary edema, which directly correlates with intracardiac pressure elevation and early sign of acute heart failure [4]. Specifically, there is a relationship between the number of B-lines seen on lung ultrasound and the amount of pulmonary congestion. There are several studies that have shown a correlation between reduction in the number of B-lines in HF patients and improvement of symptoms as well as being able to be discharged from the hospital [5, 6]. Deep-learning algorithms that can accurately assess how many B-lines are on a patient-performed lung ultrasound would support the use of ultrasound as a tele-monitoring tool for outpatients with HF. This could prove to be a quicker, cheaper, and less invasive method for following the development and prognosis of HF patients both in an inpatient and outpatient setting.