Signal Processing - July 2016 - 39

Intensity

x (Space)
α = 0.2

α = 0.5

α = 1.0

α = 1.5

α = 2.0

α = 2.5

α = 3.0

FIGURE 4. An example of the signal change magnification from a smartphone-captured video of a stable hand to obtain the heartbeat rate (based on [17]).

approach). A smartphone app therefore can use a wide range of
solutions based on goals and problem constraints. For example, a
quarantined patient can be remotely and accurately monitored by
the video magnification approach to avoid contact because power
consumption is unlikely to be a pressing issue.

Chronic wound assessment
Chronic wounds, such as pressure ulcers and diabetic wounds,
are a drain on health-care resources and, as such, continue
to challenge health-care providers to define and create more
effective intervention strategies. Wound assessment is the
foundation for maintaining and evaluating a therapeutic plan
of care. Without adequate baseline wound assessment and
valid interpretation of the assessment data, the plan of care
may be inappropriate or ineffective [19].
To reliably assess the wound type, grade (severity), and
healing process, accurate and objective measurements, such as
the area, depth, and tissue composition, are needed. Primary
measurements, including width, length, and depth, which are
critical for assessing wound severity, can be obtained through
noninvasive and visual methods. In contrast, more complex
measurements, such tunneling and undermining, are determined with invasive methods (e.g., using a cotton swab). In
addition to wound dimensions and composition, trained specialists use further information, including inflammation, temperature, exudate levels, and periwound condition (i.e., the
surrounding skin). Based on this information and years of
experience, the specialist then estimates the type and grade of
the wound and proceeds with an appropriate treatment. Therefore, it is logical to take all of these information modalities
into account for an automatic wound assessment system. The
framework that seemingly encompasses these considerations is
presented in Figure 5.

The challenge of including all of these signal modalities
is evident with regard to a smartphone app. This app should
comply with the clinical evaluation of the wound (i.e., on
what to measure and estimate) and with clinically acceptable
results to integrate into the rest of the health-care system. To
be considered an effective tool in the hands of caregivers, the
app should provide a faster and safer method than traditional
methods to assess and track wounds. Moreover, privacy issues
should be carefully examined, in addition to other on-device
processing considerations. A well-designed and accurate
smartphone-based solution for wound assessment would dramatically reduce wound-care costs and difficulties both for
patients and caregivers. Due to the capabilities of the smartphone platform, the presumed smartphone app can also be
easily integrated into the health-care ecosystem, providing
functionalities such as telemedicine and location-based emergency support systems.
Handheld solutions for wound assessment and monitoring
can be divided into two categories, with the apps in the first
category only estimating the wound size (i.e., the area and/or
volume) and the apps in the second category performing tissue classification to estimate the wound condition. We discuss
image-based mobile solutions for noninvasive wound assessment and review scientific literature and the apps markets for
Android and iOS platforms. Table 3 compares the families of
solutions, each with an example.

Wound size estimation methods
Nonautomatic methods for wound size estimation are timeconsuming and subjective and suffer from high variability.
Therefore, semiautomatic and digital planimetry techniques
have become popular and are based on the digital images
of a wound. Proposed solutions, including Wound Tracker

IEEE SIgnal ProcESSIng MagazInE

July 2016

Table of Contents for the Digital Edition of Signal Processing - July 2016