Health Diagnosis by Single Smartphone

Health Diagnosis by Single Smartphone

Lambert Spaanenburg (Lund University, Lund, Sweden)
Copyright: © 2015 |Pages: 13
DOI: 10.4018/IJHCR.2015040104
OnDemand PDF Download:
No Current Special Offers


M-Health promises to bring the ease and robustness of the thermometer to general health. Casual handling implies non-invasive measurements, where of old the hand is the main instrument to diagnose health. A smartphone or a wearable can ‘extend' the hand by embedding diagnosing intelligence. The paper discusses the role of neural modeling & matching to create such intelligence for the determination of general health factors. It is illustrated in blood pressure measurement of 90-95% medical accuracy without additional accessories.
Article Preview


Where the body is a complex physical system, there are many ways to measure and model it. The preferred way is non-invasive, as many people do not enjoy forced entry to the body. Non-invasive measurement can be based on substances available: (a) outside the body, like hairs, (b) from the skin, (c) via a cavity or (d) by fluids leaving the body. Such substances can be sampled, passed on via a strip and brought into a reader, which subsequently provides the analysis. All this remains obtrusive: one still has to take explicit action and be aware of the health issues!

Preferably health is monitored casually. That takes a small device that can be taken along during the day and that does not treat you like a potentially sick person. A fitness product is already a step in the good direction as it invigorates you as being a sporty person. However, a single mobile device is not easy to implement. The work-around is to create a detachable add-on, such as a thimble, that can be ideally fixed at the right location. The alternative _ an attachable product like a watch or an earring _ usually needs an additional phone to connect to the outside world. The advantage of the specialized part is that it is optimized to capture a specific parameter, while the disadvantage is the lack of adaptability to encompass other parameters too. This extends naturally to the “Internet of Things” (Spaanenburg, 2014).

Certain products adhere the sensor in direct contact to the skin. The electric contact to the body allows capturing electrical signals, which can be sensed, amplified and transformed for transport. There is usually more than one sensor on the same adhesive carrier, one for each health parameter of interest, which makes a miniature sensory network. The author sees such variations also in the different smart watches being developed (Poh, 2010). Products can be part of the living area, to take an active role when touched by a person. For instance, scales can communicate blood pressure, heart rate and body weight to the person’s smartphone where also temperature and body height will arrive from other sources.

The author demonstrates here that suitable models on a single smartphone can avoid external accessories. The physical process can be directly measured, but there are a number of good reasons not to do this continuously. One reason is that people do not like to be remembered to life-style sins; the other is that such measurement methods do not support high-speed sampling. The author applies the classical control-theoretic approach to learn a model, such that the observation of the modeled process is trained to the observation of the real physical process (Narendra, 1990). The connected benefit is that the model can be further studied to reveal more about the underlying process. (Figure 1). As an example the paper takes heart rate monitoring as the basis to create the necessary models for measuring blood pressure.

Figure 1.

Control theory


The paper is structured as follows. After this introduction, the model-based approach in single probe determination of general health parameters is introduced in section 2. Then the learning of the models is discussed. Further model-based blood pressure determination is presented in section 4. Finally it is concluded how beneficial neural networks are for the construction of a personal health system.


Ppg Analysis

Our health meter is of the reflective vision type. It captures the visual light reflected on capillary blood covered by skin. Capillary blood can be reached by visual light at various locations on the human body. Literature lists foot, fingertip, hand palm, pulse, tongue, ear lobe and cheek, but there is probably more. A vision sensor can capture the reflected light and the intensity shows a periodic signal in tune with the blood flow. This presence of the heartbeat reflects the presence of time in the signal, the PhotoPlethysmoGraphical or PPG signal.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing