Inter- and intra-individual variation in earprints Meijerman, L.

(1)

Citation

Meijerman, L. (2006, February 15). Inter- and intra-individual variation in earprints. Barge's Anthropologica, Leiden. Retrieved from https://hdl.handle.net/1887/4292

Version: Not Applicable (or Unknown)

License: Licence agreement concerning inclusion of doctoral thesis in the_{Institutional Repository of the University of Leiden} Downloaded from: https://hdl.handle.net/1887/4292

(2)

(3)

Abstract

(4)

8.1 Introduction

An automatic matching system for earprints may serve as a tool to extract potential matches to a query print from a database of digitised prints. Various methods towards the design of an automatic matching system for earprints have been suggested in literature (Champod et al., 2001; Ingleby et al., 2000; Rutty et al., 2005) or explored by the FearID research team (M eijerman et al., 2006; Kieckhoefer and Ingleby, 2004). Examples of explored methods are Vector Template M atching, Active Appearance M odelling and Key Point M atching. In these methods, matching algorithms provide a list of database prints that, based on the explored variables, express the greatest degree of similarity to the query print.

In order to form an opinion on individualization, an expert will compare the query print with the prints on the list of most likely matches. He not only compares similarities, but also evaluates whether differences between prints are sufficiently small as to fall within the range of potential intra-individual variation. Knowledge on the extent of possible intra-individual variation is therefore of great importance. Rendering an accurate judgment on the extent of intra-individual variation involves the examination of multiple prints of each of a great number of individuals.

This chapter is not aimed at providing an account of the limits of intra-individual variation in earprint features. Such a study should include various prints of a greater number of ears, while prints have to be taken under a greater variety of conditions known to affect the appearance of earprints. W e only aim to provide insight into the level of detail to which prints may be compared. W e show for five individuals (examples of) the variation that may occur between different prints of the same ear. To this purpose, three earprints of each included ear are compared. Similarities and differences that occur between the different prints of a single ear are described. W e further speculate on the possible causes of intra-individual variation.

8.2 M aterials and methods

(5)

listener at different sound levels. It further enabled us to record the force that was applied by the ear to the surface during the process of listening27. Each listening effort was recorded as a continuous series of consecutive force readings. Of each series – representing one listening effort – we omitted the first five and last two force readings, as these were often recorded while the subject was positioning the ear or abandoning the listening box. For each listening effort we then calculated the mean force using all remaining force readings. This value represented the functional force of a listening effort. Per listening effort, we further recorded the highest force reading of the entire series of readings.

Various sound files were played inside the box to serve as a target sound during the process of listening. They contained easy-to-answer questions, usually asked by a male voice. For subjects one to three, the level of the target sound was reduced during their second and third listening efforts as compared with their first. For subjects four and five, we used a female voice instead of a male voice during their first listening effort. We further varied the level of ambient noise. By varying conditions we aimed to increase realistic intra-individual variation in functional prints28. The duration of ear to surface contact was relatively similar for most listening efforts (17-20 seconds). For one subject one listening effort was notably longer.

The resulting earprints were dusted with aluminium powder and lifted using Black Gel Lifters. The prints were then scanned at 600dpi and subsequently imported into Adobe Photoshop software. Of a number of prints, colours were inverted to also obtain copies with a black-on-white appearance.

For each combination of prints, we described similarities and differences. Differences in position of features between the prints were emphasized by digitally overlaying a colour-inverted copy of a print on other prints of the same ear. In these overlays, those areas appearing white are representations of one print, while those in black are representations of the other. Areas common to both prints appear grey.

27

The functions of the listening box are explained in greater detail in chapters 3 (Inter- and intra-individual variation in applied force when listening at a surface, and resulting variation in earprints) and 4 (Exploring the effect of occurrence of sound on force applied by the ear when listening at a surface).

28

(6)

8.3 Results and discussion

The prints resulting from the first, second and third listening efforts are referred to as P1, P2 and P3 respectively. The calculated functional force and the recorded peak value, as well as the differences in the duration of listening between the three listening efforts, are provided in the tables for each individual (load in grams).

Force readings were made at intervals of about 0.5 seconds. The interval could in theory have varied depending on the local variation in applied force (Kieckhoefer, 2002). Consequently, we do not provide the difference in the duration of listening in seconds. The difference in the number of force readings between listening efforts may be an indication of relative differences in listening time. The recorded effort consisting of the smallest series of force readings is indicated by an ‘x’ in the data tables for each subject. A difference in the number of recordings between this and other efforts is indicated as the number of additional force readings.

Subject 1

Data P1 P2 P3

Indication of listening time x+2 readings x+3 readings x

Functional force (load in g) 1071 1439 1512

Peak value (load in g) 1752 1778 2276

1.1 Imprinted (parts of) gross anatomical features

Imprinted are: helix including part of crus of helix; stem and anterior crus of anthelix; tragus, antitragus and outline of intertragic notch; earlobe; pre- and supra-auricular areas. Mentioned features are present in all three prints.

1.2 Characteristic similarities in all three prints

(7)

antitragus-anthelix transition; 7. Longitudinal low-intensity band at superior part of helix (containing skin-detail); outer margin of helix with characteristic bulge; 8. Medial low-intensity section of anthelix in between high-low-intensity posterior margin and anterior tip; 9. Low-intensity area of helix at onset of crus of helix, inner margin angulated at this section and crus of helix section with two inwardly protruding high-intensity regions; 10. Clear pre-auricular crease anterior to tragus/crus of helix area; 11. Similar blemishes on skin of cheek; 12. Posterior auricular furrow visible; 13. Anterior notch wide, semi-square.

___________________________________________________________________________

___________________________________________________________________________ Fig. 8.1 Three prints of subject 1: P1 (a), P2 (b) and P3 (c).

___________________________________________________________________________

1.3 Apparent dissimilarities between prints

(8)

in P1 while very prominent in P3; 7. Transverse skin-lines on anthelix almost indiscernible in P1; 8. Anterior margin of anterior notch rounded in P1 and P2 while almost straight in P3; 9. Intertragic notch widest in P1; 10. Inner margin of anthelix more curved in P1.

1.4 Position of features

P1 and P3 appear to be the least similar. The position of the superior section compared with that of the inferior section is not similar in these prints. When the inferior parts of the two prints match, the superior parts of P1 are situated more anteriorly. The curvature of the anthelix is also slightly different. Fig. 8.2a provides a ‘best match’ of the crus of helix area and anterior section of the helix of both prints; in Fig. 8.2b the inferior parts and intertragic notch match.

P2 appears to be an intermediate between P1 and P3. Fig. 8.2c provides a best match of the inferior sections of P2 (inverted colours) and P3.

___________________________________________________________________________

___________________________________________________________________________ Fig. 8.2a-c Superimposed prints of subject 1: a-b, P1 (inverted colours) and P3; c, P2 (inverted colours) and P3.

(9)

1.5 Introduced variation

The tiny specs at the base of anthelix in P1 and on the inferior section of the helix in prints 2 and 3 are likely introduced through the lack of adhesion of powder to gel or of foil to gel.

1.6 Speculations on cause of variation

We suspect that changes in position and shape of the ear were due to variation in applied force. An increase in applied force appears to be affecting the superior and inferior parts of the ear differently. This is likely due to the difference in flexibility.

Applied force being relatively high for P3 would explain why the imprinted surface is generally larger in this print. Applied force being smallest for P1 could explain the occurrence of gaps in the imprinted helix. Variation in force distribution could furthermore explain differences between P1 and P2 (as well as P3). The anterior section of the helix and the anthelix appear relatively intense in P1. There may have been a greater local force on the anterior section of the helix and the anthelix, as compared with the posterior section of the helix, explaining the gaps in the latter section. Greater local force on the anthelix could also explain the reduction of the visibility of skin-lines in this print.

Differences in listening time are minimal and appear not to be expressed in the appearance of the prints.

Subject 2

Data P1 P2 P3

Indication of listening time x x+7 readings x+1 reading

(10)

present in all three prints.

1. Anterior crus of anthelix more intense than superior crus; 2. Longitudinal low-intensity band at superior section of helix, particularly towards posterior end; 3. Posterior auricular furrow visible; 4. Anthelix stem divided by fold; 5. Small transverse crease in supero-posterior corner of helix; 6. Characteristic ‘3-shaped’ inner margin of anthelix, including the small notch in between anterior crus and medial fold; 7. Earlobe demarcated from antitragus by fold (relatively obscure in P2); 8. Outline of what may be a piercing hole visible in the lobe (albeit to different extents).

___________________________________________________________________________

(11)

reduced in P2 as compared with P1, and split inferiorly in P3; 3. Interior margin of anterior section of helix less angulated in P2; 4. Tragus not extending superiorly in P1; outline of anterior notch obscured in this print; 5. Posterior auricular furrow relatively broad in P1, not continued into helix; 6. Inferior section of helix relatively narrow in P2; angle formed by the inner margin of this section and the imprinted lobe less wide; 7. Imprinted pre-auricular area relatively small in P1; supra-auricular area lacking intense pressure-area in P2.

When comparing P1 with P3 it appears that the position of features has not altered much, although it is difficult to judge as the imprinted surface is larger in the posterior section in P1, while in P3 it is larger in the anterior section. The anthelix appears somewhat more ‘upright’ in P1; it appears impossible to match the sections superior and inferior to the medial fold at the same time. Fig. 8.4a shows a best match of the intertragic notch and inner margin of the helix of P1 and P3.

The helix is shaped quite differently in P2 as compared with P1 and P3. Fig. 8.4b shows a best match of the intertragic notches of P2 and P3, revealing a slight change of position of the anthelix. Fig. 8.4c shows a best match of the anthelix.

The square on the lobe in P1 appears unnatural; the thin lines on antitragus, lobe and tragus in all prints may be imprinted hairs. The (additional) blurred surface between the anthelix stem and the inferior section of the helix could be due to movement of the ear during the process of listening, but could also be due to brush strokes. Differences in the fine details of the superior crus of the anthelix appear to be due to a variation in the direction of brush strokes.

(12)

helix shape is probably due to the difference in force or force distribution. A difference in force distribution is likely also the cause of the difference between P1 and P3 in the imprinted surface of the anterior and posterior regions.

__________________________________________________

___________________________________________________________________________ Fig. 8.4a-c Superimposed prints of subject 2: a, P3 (inverted colours) and P1; b-c, P3 (inverted colours) and P2.

___________________________________________________________________________

Subject 3

Data P1 P2 P3

Indication of listening time x+31 readings x+1 reading x

(13)

tragus, antitragus and outline of intertragic notch; earlobe; pre-auricular area. Mentioned features are present in all three prints.

1. Inferior section of helix more prominently imprinted than superior section; 2. Fold at apex of intertragic notch; 3. Pattern of pre-auricular creases (the very short branch diverting from the anterior crease is very characteristic); 4. Transverse and longitudinal creases in antitragus; 5. Characteristic double crease extending from lacuna in lobe; 6. Trace of possible auricular tubercle; 7. Anterior notch pointed; inferior margin somewhat curved; 8. Small isolated pressure-area in anterior region of helix.

___________________________________________________________________________

(14)

anthelix notched at this position in P1 and P2; 3. Imprinted surface of antitragus largest in P1; smallest in P2; 4. Crease extending from apex of scapha not visible in P1; 5. Posterior section of helix relatively narrow in P1; 6. Area antero-inferior to intertragic notch more intense than remaining parts of tragus in P1 and P3 but not in P2.

A comparison of P1 and P2 shows that the posterior section of the print, i.e., the helix, is situated more posteriorly in P1: the ear appears wider. Fig. 8.6a shows a best match of area surrounding the intertragic notch in P1 and P2.

P2 and P3 appear more similar. The anthelix is, however, situated more posteriorly in P2. The pre-auricular creases appear to be similar in position to one other, but not to the imprinted ear. Fig. 8.6b provides a best overall match of the imprinted areas. In Fig. 8.6c, the pre-auricular creases match, illustrating the difference in position of the ear as compared with the pre-auricular area between these two prints.

___________________________________________________________________________

__________________________________________________________________________ Fig. 8.6a-c Superimposed prints of subject 3: a, P1 and P2 (inverted colours); b-c, P2 (inverted colours) and P3.

(15)

The squared lacunae in the pre-auricular area of P1 appear to be due to a lack of adhesion of the powder.

The ear appears wider in P1. Total applied force was relatively low during the first listening effort. This is surprising, as one would expect the ear to become wider (flatter) when applying more force. There may, in theory, have been a discrepancy between the difference in total force and the force applied by the ear and by the pre-auricular area respectively, but we see no evidence of this in the pre-auricular area.

A difference in pressure distribution across the ear could explain the reduction in imprinted helix surface, in combination with the increase in anthelix surface, in P1 as compared with P2 and P3. The duration of ear-surface contact was relatively long for P1, which could also underlie the larger anthelix and lobe surface.

Subject 4

Data P1 P2 P3

Indication of listening time x+8 readings x x+7 readings

Imprinted are: helix including part of crus of helix; stem and anterior crus and trace of posterior crus of anthelix; tragus, antitragus and outline of intertragic notch; earlobe; pre- and supra-auricular areas. Mentioned features are present in all three prints.

(16)

tubercle; 9. Supero-posterior section of helix narrow and low in intensity.

___________________________________________________________________________

1. Additional surface in posterior section of helix in P2; 2. Inferior section of helix relatively narrow (inferior section of scapha broad) in P2; 3. Anterior notch more demarcated in P1; 4. Void between helix and pre-auricular area relatively small in P1 while wide in P3.

(17)

__________________________________________________________________________________________

___________________________________________________________________________ Fig. 8.8a-c Superimposed prints of subject 4: a-b, P1 (inverted colours) and P3; c, P1 (inverted colours) and P2.

___________________________________________________________________________

The horizontal line across the pre-auricular area in P3, and the dark patch in the superior section of the helix in P2, are likely introduced. A difference in the direction of powdering caused a difference in the direction of the thin lines in the anterior (low-intensity) section of the pre-auricular area.

(18)

Subject 5

Data P1 P2 P3

Indication of listening time x+3 readings x x+1 reading

Imprinted are: helix including a trace of crus of helix; stem and anterior crus of anthelix, as well as a minimal trace of posterior crus; tragus; antitragus; traces of earlobe; pre-auricular area. Mentioned features are present in all three prints.

___________________________________________________________________________

(19)

section (less obvious in P1); 3. Creases and voids anterior to tragus; 4. Interruption of helix in supero-posterior section; 5. Pressure-area anterior to helix near onset of crus of helix; 6. Void between anthelix and antitragus; 7. Transverse skin lines on anthelix (superior and inferior sections) and oblique crease on anthelix stem; 8. Interruption of earlobe margin (due to earring?); 9. Two small notches in inferior margin of anterior crus of anthelix; 10. Similar ‘meandering’ of inner margin in anterior section of imprinted ear, from angle and minute notches in helix margin to notch in tragus margin; 11. Low-intensity patches and tiny voids near outer margin of antero-superior section of helix; 12. Small pressure-area in lacuna between lobe and antitragus.

1. Anterior section of helix less well demarcated from pre-auricular area in P1; 2. Void between anthelix and antitragus larger in P1; 3. Void between tragus and pre-auricular area split in P2; 4. Intertragic notch relatively wide in P1, and narrow in P2; 5. Imprinted area of lobe larger in P2; 6. Notch at apex of scapha in P2; 7. Anthelix stem relatively narrow in P2; broadest in P3; 8. Inferior section of helix relatively broad in P2.

When the anthelix and the superior section of the helix of P1 and P2 match, the tragus and antitragus do not. In Fig. 8.10a, the helix and anthelix match. In P1, the inferior parts of the helix and anthelix appear to be situated slightly more anteriorly. In addition, the ear appears slightly wider inferiorly.

P1 and P3 appear relatively similar with respect to the position of features. The anthelix and superior parts of the helix are curved slightly different and the ear again appears slightly wider in P1 (see Fig. 8.10b).

P2 and P3 appear least alike. When the anthelix matches relatively well, the tragus and antitragus do not (Fig. 8.10c).

(20)

___________________________________________________________________________

___________________________________________________________________________ Fig. 8.10a-c Superimposed prints of subject 5: a, P1 (inverted colours) and P2; b, P1 (inverted colours) and P3; c, P2 (inverted colours) and P3.

___________________________________________________________________________

The pressure distribution was perhaps different for all three prints: higher pressure on the anterior section of the helix during the first listening effort (helix not well demarcated from pre-auricular region), a relatively high amount of pressure on the inferior section of helix during the second effort (broader helix, more lobe imprinted, narrower anthelix stem), and a relatively high amount of pressure on the anthelix during the third (more intense and broader anthelix, narrow helix).

8.4 Conclusion

(21)

No golden rule may be provided here to distinguish between inter- and intra-individual variation. With respect to applied force, for instance, it was found that a relatively similar force variation between listening efforts may cause different variation in the resulting earprints. Pressure distribution appears to be an important factor in determining the appearance of imprinted features. We noticed no trend towards a loss of imprinted detail that could have been attributed to the reduction of oily substances on the ear during consecutive efforts of listening.

The observed differences between prints of the same ear appear small as compared with the differences among prints of the five individuals. Whether or not intra-individual differences are sufficiently small as to be able to group prints of the same ear together in a sample of prints that expresses a relatively small amount of inter-individual differences will be addressed in chapter 9.

Acknowledgements