A hemoglobin concentration dataset derived from triple-wavelength photoplethysmography for machine learning applications

The commonly used method for measuring Hemoglobin (Hb) concentration involves invasive blood sampling, which has drawbacks such as discomfort and pain. To address these limitations, photoplethysmography (PPG) has emerged as a promising non-invasive alternative. PPG device is a non-invasive optical device used to detect blood volume changes within microvascular tissue beds. This device offers the capability to measure hemoglobin levels without the need for invasive procedures. In this article, the dataset consists of PPG signal, Hb value, age, and gender. The PPG signal was measured by a MAXM86161 module sensor that emitted triple-wavelength light (green, red and infrared). Data was collected from 52 subjects (60 % female, 40 % male), with Hb values ranging from 12 to 20 g/dL for males and 8-18 g/dL for females. Acquisition occurred from April to May 2025 at West Java, Indonesia. The public dataset on Mendeley Data includes raw PPG signals, filtered outputs, and peak detection results for each subject. This resource is valuable for researchers advancing non-invasive Hb measurement based on machine learning.