Non-invasive anaemia detection by examining palm pallor: A smartphone-based approach. (January 2023)

• Record Type: Journal Article
• Authors: Kesarwani, Abhishek; Das, Sunanda; Dalui, Mamata; Kisku, Dakshina Ranjan; Sen, Bibhash; Roy, Suchismita; Basu, Anupam
• Is Part Of: Biomedical signal processing and control. Volume 79(2023)Part 1
• Publication Date: 2023-01
• Subjects: Anaemia -- Palm pallor -- Image registration -- ROI detection -- Prediction models; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.104045
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract: Anaemia, caused due to lack of blood haemoglobin level, is one of the most common diseases that affects billions of people across the world. According to WHO statistics, India is one of the developing countries with highest prevalence of anaemia. Conventional invasive methods are cost-prohibitive and difficult to administer globally which essentially demands non-invasive, inexpensive and reliable solution. This work targets to combine state-of-the computational approaches with the idea of age-old practice of rough estimation of blood haemoglobin level by observing pallor in palm to develop a non-invasive anaemia detection system. The proposed system works with the basis of inducing changes in palm pallor with suitable pressure application and release, measuring the rate of change of colour and performing time-domain analysis thereof to correlate with blood haemoglobin concentration. The video of colour changes in palm induced through a customised device, is captured using smartphone camera and is processed and analysed through a set of image processing and analysis techniques. Estimated parameters are provided as input to the ensemble-based prediction models comprising of regression and multi-layer perceptron. The proposed system ensures a sensitivity of 93% with MSE and RMSE of 0.701 and 0.698, respectively, determined on palm pallor video samples of 131 participants.