Page 67 - Summer 2021

Figure 6. Medical ultrasound classification of cancerous tissue using echo statistics parameters (k and μ) in combination with acoustically inferred scatterer size (Effective Scatterer Diameter) to classify different types of cancer. The terms k and μ are from the generic statistical function, the homodyned K-distribution, and were determined from a best fit of that function to echo statistics data (i.e., the histogram). Adapted from Oelze and Mamou, 2016, with permission of the IEEE.
of cancer (Oelze and Mamou, 2016). In this case, the variability in echo level is a function of the variability of the material properties (sound speed and density) of the tissue, and those properties are, in turn, dependent on tissue type. As an aid in discrimination, it has been found to be useful to include an acoustic inference of scatterer size within the tissue. The size is estimated by compar- ing predictions based on scattering models and the data. This estimate has been combined with two parameters of the echo statistics for classification and discrimination between different types of tissue (Figure 6).
Echolocating Mammals: Dolphins and Bats
It is currently impossible to state definitively how dolphins, bats, and other echolocating mammals “process” their echo data. However, studies demonstrate that the mam- mals are using sound to discriminate among scatterers with different echo signatures to identify their prey (Sim- mons, 2017; Tyack, 2017). Because it is known that prey with different anatomies have different acoustic signatures, it is reasonable to assume that the mammals are, in some way, using the differences in echo characteristics as a basis for discrimination between those different types of prey. For example, the echoes from moths (a prey of bats) with
different anatomies have been shown to possess different modulations from the fluttering of their wings and, hence, different echo statistics (Lee and Moss, 2016). In addition, a Blainville beaked whale has been observed to target (and eat) a type of scatterer (presumed prey) that possessed a particular echo-frequency spectrum (Jones et al., 2008).
Beyond Acoustics: Radars and Lasers
The principles of echo classification extend well beyond acoustics to important applications involving radars and lasers (Goodman, 1985; Watts and Ward, 2010). After all, whether it be acoustics (a mechanical vibration) or radar/ laser (an electromagnetic phenomenon), the signals all travel as a wave that has properties including scattering and interference (Jakeman and Ridley, 2006). For exam- ple, radars are routinely used in air traffic control on the ground to track incoming and outgoing aircraft. They are also used on ships to aid in navigation and detect unknown nearby vessels.
Radar Classification of Ships
In one controlled study, a radar mounted on a satellite was used to detect ships and oil rigs on the ocean (Ferrara et al., 2011). The data contained a complex superposition of echoes from the sea surface and the various metallic struc- tures that were on the surface. In some cases, it was difficult to distinguish between (1) the occasional high values of the
Figure 7. Radar detection and classification of ships using a satellite-deployed radar. Echo statistics parameters are used to separate radar echoes from ships and the sea surface. Center and right: raw and processed echoes, respectively, arranged in an aerial view of the data. T1-T7 are echoes from ships. Center: grainy background echoes are from the sea surface. Adapted from Ferrara et al., 2011, with permission of the IEEE.
Summer 2021 • Acoustics Today 67