Measuring Heart Rate from Video. Bush, I. & Mall, S. abstract bibtex A non-contact means of measuring heart rate could be beneficial for sensitive populations, and the ability to calculate pulse using a simple webcam or phone camera could be useful in telemedicine. Previous studies have shown that heart rate may be measured in color video of a person’s face. This paper discusses the reimplementation of one such approach that uses independent component analysis on mean pixel color values within a region of interest (ROI) about the face. We explore the idea further by assessing the algorithm’s robustness to subject movement and bounding box noise and examine new means of choosing the ROI, including segmentation of facial pixels through a reimplementation of GrabCut. Heart rate was measured with an error of 3.4 ± 0.6 bpm in still video and 2.0 ± 1.6 bpm in video with movement. Facial segmentation improved the robustness of the algorithm to bounding box noise.
@article{bush_measuring_nodate,
title = {Measuring {Heart} {Rate} from {Video}},
abstract = {A non-contact means of measuring heart rate could be beneficial for sensitive populations, and the ability to calculate pulse using a simple webcam or phone camera could be useful in telemedicine. Previous studies have shown that heart rate may be measured in color video of a person’s face. This paper discusses the reimplementation of one such approach that uses independent component analysis on mean pixel color values within a region of interest (ROI) about the face. We explore the idea further by assessing the algorithm’s robustness to subject movement and bounding box noise and examine new means of choosing the ROI, including segmentation of facial pixels through a reimplementation of GrabCut. Heart rate was measured with an error of 3.4 ± 0.6 bpm in still video and 2.0 ± 1.6 bpm in video with movement. Facial segmentation improved the robustness of the algorithm to bounding box noise.},
language = {en},
author = {Bush, Isabel and Mall, Serra},
pages = {8},
}
Downloads: 0
{"_id":"xk4rsmHttBjpP5cAX","bibbaseid":"bush-mall-measuringheartratefromvideo","author_short":["Bush, I.","Mall, S."],"bibdata":{"bibtype":"article","type":"article","title":"Measuring Heart Rate from Video","abstract":"A non-contact means of measuring heart rate could be beneficial for sensitive populations, and the ability to calculate pulse using a simple webcam or phone camera could be useful in telemedicine. Previous studies have shown that heart rate may be measured in color video of a person’s face. This paper discusses the reimplementation of one such approach that uses independent component analysis on mean pixel color values within a region of interest (ROI) about the face. We explore the idea further by assessing the algorithm’s robustness to subject movement and bounding box noise and examine new means of choosing the ROI, including segmentation of facial pixels through a reimplementation of GrabCut. Heart rate was measured with an error of 3.4 ± 0.6 bpm in still video and 2.0 ± 1.6 bpm in video with movement. Facial segmentation improved the robustness of the algorithm to bounding box noise.","language":"en","author":[{"propositions":[],"lastnames":["Bush"],"firstnames":["Isabel"],"suffixes":[]},{"propositions":[],"lastnames":["Mall"],"firstnames":["Serra"],"suffixes":[]}],"pages":"8","bibtex":"@article{bush_measuring_nodate,\n\ttitle = {Measuring {Heart} {Rate} from {Video}},\n\tabstract = {A non-contact means of measuring heart rate could be beneficial for sensitive populations, and the ability to calculate pulse using a simple webcam or phone camera could be useful in telemedicine. Previous studies have shown that heart rate may be measured in color video of a person’s face. This paper discusses the reimplementation of one such approach that uses independent component analysis on mean pixel color values within a region of interest (ROI) about the face. We explore the idea further by assessing the algorithm’s robustness to subject movement and bounding box noise and examine new means of choosing the ROI, including segmentation of facial pixels through a reimplementation of GrabCut. Heart rate was measured with an error of 3.4 ± 0.6 bpm in still video and 2.0 ± 1.6 bpm in video with movement. Facial segmentation improved the robustness of the algorithm to bounding box noise.},\n\tlanguage = {en},\n\tauthor = {Bush, Isabel and Mall, Serra},\n\tpages = {8},\n}\n\n","author_short":["Bush, I.","Mall, S."],"key":"bush_measuring_nodate","id":"bush_measuring_nodate","bibbaseid":"bush-mall-measuringheartratefromvideo","role":"author","urls":{},"metadata":{"authorlinks":{}},"html":""},"bibtype":"article","biburl":"https://bibbase.org/zotero/shamimown","dataSources":["Wo3KRwzN9T9tb8ucf"],"keywords":[],"search_terms":["measuring","heart","rate","video","bush","mall"],"title":"Measuring Heart Rate from Video","year":null}