Abstract
In many countries, bare footprints collected from crime scenes can be used as evidence for forensic identification, which involves the linear measurement of quantitative characteristics. Compared with several other methods of footprint measurement, the Reel method has been proven to be reliable and has been used in many countries. The Chinese method for bare footprint measurement is the official method issued by Ministry of Public Security of China, but its reliability has not been tested. This study focuses on the reliability test of the Chinese method for bare footprint measurement, which can serve as a reference for the verification in the process of footprint identification. Sixty-four volunteers were randomly selected to make dust and inked footprints, and five metrics of each footprint were measured and re-measured by three raters to test the reliability. These metrics consisted of four main ones from the Chinese method, with one from the Reel method for comparison. Based on intraclass correlation coefficients, the standard error of measurement and 95% Bland–Altman limits of agreement, test–retest reliability, and rater reliability within and among three raters were analysed. The Chinese method finally demonstrated a high degree of reliability. Although the Reel method achieved a marginally higher reliability score, its advantage was slight. Furthermore, the Chinese method for bare footprint measurement is not only reliable but also simple to operate, making it a viable supplement to the Reel method in contexts outside of China. In the process of footprint identification in China, when footprints are re-measured or measured by different appraisers for verification, the reliability values given in this paper could be used as a reference.
Keywords: forensic sciences, footprint, trace inspection, linear measurement, reliability test
Introduction
In some criminal cases, bare footprints can be obtained from the crime scene and used as suspicious prints in forensic identification. The comparison of suspicious prints and known prints allows for a conclusion on whether they were made by the same suspect, and therefore can be used as evidence in court. Different from fingerprints, bare footprints extracted from crime scenes often show few ridge features [1], and the similarity evaluation mainly depends on category features such as shape or size, which can be transformed into descriptive features and quantitative features. The combination of category characteristics and sub-category characteristics can show strong specificity [2]. There are various methods to measure quantitative characteristics, such as the Gunn method, the Kennedy method, and the Reel method [2–5]. However, in China, there are official standards for footprint identification, issued by the Ministry of Public Security of China, which involve the examination of morphological features and linear features [6–8]. Therefore, a Chinese method for the linear measurement of footprints exists for forensic use.
Different from other countries, the identification of bare footprints involved in criminal cases in China is completely carried out by the criminal technical department of the police, and the measurement methods involved in the identification should be carried out in accordance with the standards issued by the Ministry of Public Security of China. Since the late 1950s, bare footprint evidence has been gradually applied in court trials. In this paper, China’s standard method for measuring barefoot footprints is termed the “Chinese method”.
In China, the methods and standards for barefoot footprint inspection and identification originated from the training and education provided by criminologists from the former Soviet Union in the 1950s and have been revised and improved since then [9, 10]. In the identification of bare footprints, the reference line chosen for quantitative measurement of footprints generally adopts the connecting line between the center of the second toe of the foot and the midpoint of the heel area, which is different from other methods mentioned above. After 2017, the Ministry of Public Security of China organized experts to draft the relevant standards for barefoot footprint identification [6–8], and set the reference line for footprint measurement as the connecting line between the center of the second toe and the outermost point of the heel (when the second toe is not displayed, the midpoint of the connecting line between the center of the first toe and the center of the third toe is selected as the reference point). Based on the reference line, four main linear features can be measured in the barefoot footprint identification process.
The validity and reliability test of the method for footprint measurement is important, which is the first step in the basic study of bare footprint identification [5]. Reel [11] used a sample of 30 footprints obtained with an inkless kit to conduct a metric analysis on the reliability of four methods (the Gunn Method, the Optical Center Method, the Kennedy method, and the Reel method). She obtained a reliability score which was important for comparing linear measurements. Nirenberg et al. [12] also selected a footprint sample of 30 participants to prove the Reel method is reliable for the linear measurement of sock-clad footprints. By defining a stable baseline, the Reel method demonstrated higher stability than the other three methods. Based on a reliable measurement method, the procedure of bare footprint identification can be reliable and explicable for forensic use [13].
In several countries such as the USA, UK, and India, the Reel method has been verified as reliable and has been used not only in forensic identification but also in the clinical and anthropological fields [13–21]. In China, the Chinese method for footprint measurement is used as the official standard for barefoot footprint identification, but its reliability has not been formally verified yet.
The primary aim of this study was to evaluate the validity and reliability of the Chinese method for bare footprint measurement, based on statistical research. In addition, although new software was used for measurement in this study, the reliability of the Reel method was also tested for comparison. This study collected both inked and dust footprints to assess measurement reliability. Most previous studies have used high-quality samples, such as inked footprints or those obtained from inkless paper systems [3–5, 12, 13,15–24]. By contrast, footprints encountered at crime scenes are often dust impressions which are typically of lower quality. Therefore, a sample set encompassing both high and low qualities may be more representative of actual casework, potentially yielding more comprehensive and applicable results.
Materials and methods
Since the sample set included both dust footprints and inked footprints, more participants were recruited for this study than in previous ones [12,13] to test the reliability of the measurement method. In this study, 64 adults (54 males and 10 females) of the same ethnicity were randomly recruited from a college to collect bare footprints. They are all in good health and have no foot abnormalities or abnormal walking habits. Their ages range from 22 to 24 years, with a small variation. All participants provided written informed consent. Furthermore, standardized instructions were given for the footprint collection process, which all participants followed rigorously.
Collection of footprints
The 64 participants were evenly divided into two groups of 32. One group provided inked prints, and the other provided dust prints. For each group, half of the participants left static footprints and the other half left dynamic footprints. Before sample collection, each participant was asked to wash their feet first and wipe them up afterwards in order to make their footprints unpolluted.
For the first group, 16 participants were instructed to provide static footprints, and the other 16 participants provided dynamic footprints. To collect static footprints, participants were first instructed to put their feet onto the inked pad. They then stepped onto the paper with both feet, bearing their full weight for a while to leave the static footprints. Eight participants were asked to leave their left footprints, and the other eight left their right footprints. The participants who leave dynamic footprints were asked to follow the mid-protocol [1]. They walked normally, stepped their right foot onto the inked pad, and kept on walking to leave their dynamic right footprint onto the paper two steps ahead. For the convenience of dynamic footprint sample collection, only the right footprints were collected. After collection, all 32 footprints (16 static and 16 dynamic) were scanned using an HP scanner with 100 dpi to convert them into electronic images for subsequent measurements.
For the other group, a similar procedure was used to collect the static and dynamic dust footprints, which involved 16 static prints and 16 dynamic prints. Just like the inked sample set, the dust footprint sample included 16 footprints of the right foot and 16 footprints of the left side. Participants were asked to put their feet into a dust pad for a while and then stand or walk on a hard tile floor to leave footprints. To obtain electronic images of dust footprints, a portable footprint scanner provided by Hangzhou Chuangheng Electronic Technology Development Co., Ltd (Hangzhou, China) was used (Figure 1), which can automatically adjust the lighting angle to capture images at their original scale without distortion. The footprints were scanned using this device at a resolution of 200 dpi.
Figure 1.
(A) Footprint scanner and (B) one of its scanned photos (dust footprint).
Table 1 shows the specifics of the footprints grouping. Finally, 64 electronic bare footprints were collected from the two groups. To better preserve the original information of the images, the stored footprint images are in BMP format. Each of the groups contains footprints from both sides, with half of the footprints in the static state and the other half in the dynamic state. To reach a persuasive conclusion, a sample with more heterogeneity could be more representative. During the footprint collection, any prints that happened to be smudged or too faint were excluded and collected again.
Table 1.
Specifics of the footprint sample grouping.
| Static (right) | Static (left) | Dynamic (right) | Dynamic (left) | Total | |
|---|---|---|---|---|---|
| Dust footprints | 8 | 8 | 8 | 8 | 32 |
| Inked footprints | 8 | 8 | 8 | 8 | 32 |
Measurement of linear metrics
For the convenience of measurement, a multifunctional software named IC Measure (version 2.0.0.286, https://www.theimagingsource.com/en-us/support/download) was used for the measurements. This measurement software was found through the internet search engine. Compared with other free measuring software such as Photoshop or GIMP described in previous studies [1,12,13], IC Measure is easy to manipulate and has many convenient functions for measurements. Similar to Photoshop and GIMP software, IC Measure is an image analysis tool based on pixel measurement. It uses fewer system resources, is available in multiple versions, and can be installed on computers running Windows XP or later with at least 2G of memory. In China, the requirement for measurement accuracy in footprint examination is 0.1 cm. According to the minimum measurement accuracy of IC Measure, this requirement can be fully met.
As described in the Chinese method for footprint identification [6], bare footprints features comprise both morphological characteristics (which are categorical) and linear measurements (which are quantitative). This metric study employs the linear measurements. This process involves confirming the mid-line (reference line) and includes the four main measures listed below (Figure 2).
Figure 2.
The Chinese method for bare footprint measurement. Mid-line of the footprint: a line that connects the central point of the second toe and the outermost of the heel. (When the second toe is not displayed, the midpoint of the connecting line between the center of the first toe and the center of the third toe is selected as the reference point). Overall length of the footprint (Overall length): the distance between two lines which is perpendicular to the mid-line and tangent to the outermost of the toes and the heel. Ball width of the footprint (Ball width): the distance between two lines which is parallel to the mid-line and tangent to the inner and outer edges of the foot sole. Arch width of the footprint (Arch width): the length of the line which is perpendicular to the mid-line of the footprint and comes through the narrowest part of the arch area. Heel width of the footprint (Heel width): the length of the line which is perpendicular to the mid-line and comes through the center of the heel.
While a precise Mid-line of the footprint is critical to reach a reliable measurement result, the “circle tool” of IC Measure was used to confirm the central point of the second toe.
Three raters were randomly selected to measure the group of inked footprints. All these raters graduated from a police academy where they majored in forensic science, and each had >3 years of experience in crime scene investigation, trace identification, and the examination of footprints. In this study, they used the same type of computers (Windows 10 system version) for measurements, with a screen resolution of 1 366 × 768. These raters were asked to measure these footprints for the evaluation of rater reliability (an inter-rater study). In addition, Rater 1 measured these metrics two times for the evaluation of test–retest reliability (an intra-rater study). For the rigour of the study, each rater was asked to measure these metrics separately and independently. To eliminate the influence of fatigue, raters who felt tired during the measure were asked to rest at least 10 min for the next measure, and the Rater 1 had a time interval of >24 h before re-measurement. As the ghosting areas may appear in the dynamic footprints [22,23], they were all included in the measurement.
In order to make a comparison of the Chinese method to the Reel method, the Calc_A1 measure of the Reel method [5] was also measured and analysed for each of these footprints. The operation manual of the Reel method was provided first and followed strictly during the measurement. To establish a precise central axis for the Reel method, the angle tool of IC Measure was used to define the central axis, and Calc_A1 was then measured from all prints. To ensure the rigour of the reliability study, all central axes were re-confirmed before re-measurements. Finally, five metrics, which included one metric from the Reel method, were measured from each footprint.
For the dust footprint sample, Rater 1 was assigned to assess test–retest reliability, while Rater 2 and Rater 3 were involved in the assessment of rater reliability. Similar to the procedure for the inked footprints, four metrics from the Chinese method and the Calc_A1 metric from the Reel method were collected from all dust footprints.
Statistical analysis
As a reference of statistical analysis methods used in clinic and health research [24–27], and the former reliability study carried out by Sarah Reel et al. [5,12], intraclass correlation coefficient (ICC), standard error of measurement (SEM) and Bland-Altman plots with 95% confidence intervals for the limits of agreement (95%CI of LOA) models were used to assess the reliability of the Chinese method. ICC was used to evaluate the relative reliability, while SEM and LOA were used to assess the absolute reliability [12]. The measured data from dust and inked footprints were recorded separately as independent groups. Then test–retest reliability and rater reliability were assessed in each group. For each group, the five metrics of each footprint were assessed by statistical analysis applications.
After all measurement data were entered into the computer, they were prepared for ICC calculation using SPSS software (version 23, IBM Corporation, Armonk, NY, USA). A one-way random effects model ICC was applied to assess test–retest reliability test, whereas a two-way random effects model ICC was used for the rater reliability tests. SEM was calculated using the equation: 
. At last, MedCalc software (version 22, https://www.medcalc.org) was used to plot the 95%CI of LOA.
Results
Each metric measured in the subgroups showed good normality through Shapiro–Wilk tests (P > 0.05) and Q-Q plots. Therefore, parametric tests were appropriate for all these reliability tests [25].
Reliability test results of inked footprints
Test–retest reliability results
Table 2 shows the ICC, SEM, and the 95%CI of LOA of test–retest reliability, while Figure 3 shows the Bland–Altman plot of LOA. ICC demonstrates a high score between 0.989–0.999, and SEM floats from 0.025 to 0.091 cm. On average of these metrics, the SEM that demonstrates the standard error between measured value and true value is less than 0.1 cm. The 95%CI of LOA of these metrics shows a short length of <0.5 cm, among which the Arch width had the longest interval and the Heel width had the shortest interval.
Table 2.
Intra-rater reliability analysis of selected metrics from Chinese method and Reel method (inked footprints).
| Measurements | ICC | SEM (cm) | 95%CI of LOA (cm) |
|---|---|---|---|
| Overall length | 0.999 | 0.047 | −0.1387, 0.0812 |
| Ball width | 0.995 | 0.043 | −0.1420, 0.0877 |
| Arch width | 0.989 | 0.091 | −0.2189, 0.2820 |
| Heel width | 0.997 | 0.025 | −0.0741, 0.0428 |
| Calc_A1 (Reel method) | 0.999 | 0.052 | −0.1438, 0.0788 |
ICC: intraclass correlation coefficient; SEM: standard error of measurement; LOA: limit of agreement.
Figure 3.
Plots of Bland-Altman plots with 95% confidence intervals for the limits of agreement (95%CI of LOA) of test–retest measurements (inked footprints).
Rater reliability results
Table 3 shows ICC and SEM among all three raters, and Figure 4 shows 95%CI of LOA by Rater 2 and Rater 3. What stands out is the result of the Ball width, which had the lowest ICC and largest SEM. In the rater reliability test, Calc_A1 from the Reel method got the best reliability result compared to the metrics from the Chinese method.
Table 3.
Inter-rater reliability analysis of selected metrics from Chinese method and Reel method (inked footprints).
| Measurements | ICC (R1, R2, R3) | SEM (cm) (R1, R2, R3) | 95%CI of LOA (cm) (R2 and R3) |
|---|---|---|---|
| Overall length | 0.996 | 0.094 | −0.1078, 0.2253 |
| Ball width | 0.905 | 0.184 | −0.5849, 0.0126 |
| Arch width | 0.978 | 0.128 | −0.4026, 0.2082 |
| Heel width | 0.984 | 0.059 | −0.1443, 0.2249 |
| Calc_A1 (Reel method) | 0.999 | 0.051 | −0.1436, 0.0512 |
R1, R2, R3: Rater 1, Rater 2, Rater 3; ICC: intraclass correlation coefficient; SEM: standard error of measurement; 95%CI: 95%confidence interval; LOA: limit of agreement.
Figure 4.
Plots of Bland-Altman plots with 95% confidence intervals for the limits of agreement (95%CI of LOA) of inter-raters measurements (inked footprints).
Reliability test results of dust footprints
Test–retest reliability results
As shown in Table 4 and Figure 5, the test–retest reliability results for the dust footprints are presented. All metrics demonstrate high relative reliability, with ICC values ranging from 0.980 to 0.999 and SEM values were all below 0.11 cm. In contrast to the results from the inked footprints, the Ball width did not show higher reliability than the Arch width, and exhibited the largest 95%CI of LOA among these metrics.
Table 4.
Intra-rater reliability analysis of selected metrics from Chinese method and Reel method (dust footprints).
| Measurements | ICC | SEM (cm) | 95%CI of LOA (cm) |
|---|---|---|---|
| Overall length | 0.996 | 0.083 | −0.2400, 0.1937 |
| Ball width | 0.980 | 0.099 | −0.2501, 0.2988 |
| Arch width | 0.994 | 0.103 | −0.1891, 0.3266 |
| Heel width | 0.992 | 0.049 | −0.1502, 0.1090 |
| Calc_A1 (Reel method) | 0.999 | 0.042 | −0.1123, 0.1410 |
ICC: intraclass correlation coefficient; SEM: standard error of measurement; 95%CI: 95%confidence interval; LOA: limit of agreement.
Figure 5.
Plots of Bland-Altman plots with 95% confidence intervals for the limits of agreement (95%CI of LOA) of test–retest measurements (dust footprints).
Discussion
This study evaluated the validity and reliability of the Chinese method for measuring bare footprints and compared it with Sarah Reel’s linear measurement method. IC Measure software was chosen for the measurement of footprint photographs, which is different from the Sarah Reel’s recommendation to use GIMP software. The Reel method has been proven to demonstrate high reliability when implemented with GIMP software, and it is more widely adopted than other methods, such as the Gunn, Kennedy and Optical Centre methods. Therefore, only the Reel method is selected for comparison with the Chinese method. As the present study incorporated both dust and inked footprints, approximately twice the number of volunteers were recruited compared to Reel’s study to generate the footprint samples. A recognized limitation of this study is the relatively homogeneous age and ethnic background of the participant group. However, in the actual measurement process, the effect of factors such as age and sex on the footprints was mainly reflected in differences in the length of each part, which did not significantly affect the measurement results. Comparatively, factors such as left or right foot and different medium conditions (dust or ink, etc.), can alter the image morphology and quality, which may have a greater impact on the measurements. Therefore, to account for this potential heterogeneity, the left and right foot, along with different medium conditions, was controlled as key variables in this study.
A comparison of the reliability of the Reel method with that reported in the previous study [12] also confirms the reliability of the IC Measure. The results indicate that IC Measure also achieves high reliability, comparable to that of GIMP software. For example, the intra-rater ICC for Calc_A1 from the Reel method measured on inked footprint was 0.999, consistent with the result reported by Sarah Reel [12]. More important than the choice of software, careful measurement practices are crucial for obtaining highly reliable results.
Rater reliability results
Table 5 and Figure 6 show the rater reliability results for the metrics obtained from the dust footprints. Overall, the distribution of these results was similar to those observed for the inked footprints. It can be observed that the Calc_A1 from the Reel method achieved the highest ICC. For the Chinese method, the Ball width yielded the lowest ICC, while the Heel width demonstrated the highest reliability.
Table 5.
Inter-rater reliability analysis of selected metrics from Chinese method and Reel method (dust footprints).
| Measurements | ICC | SEM (cm) | 95%CI of LOA (cm) |
|---|---|---|---|
| Overall length | 0.989 | 0.138 | −0.4380, 0.2767 |
| Ball width | 0.917 | 0.210 | −0.5211, 0.6449 |
| Arch width | 0.968 | 0.238 | −0.4885, 0.7517 |
| Heel width | 0.968 | 0.100 | −0.2313, 0.3157 |
| Calc_A1 (Reel method) | 0.990 | 0.131 | −0.3983, 0.3402 |
ICC: intraclass correlation coefficient; SEM: standard error of measurement; 95%CI: 95%confidence interval; LOA: limit of agreement.
Figure 6.
Plots of Bland-Altman plots with 95% confidence intervals for the limits of agreement (95%CI of LOA) of inter-raters measurements (dust footprints).
Comparison between inked prints and dust prints for measurement reliability
Forensic analysis of bare footprints relies on samples collected from crime scenes, which are typically dust-based and of low quality. Previous studies on footprint reliability tests have predominantly focused on inked prints, often overlooking dust footprints. To address this gap and ensure methodological rigor, both dust and inked prints were included in the present study.
The ICC results showed high relative reliability for both dust and inked footprints, confirming that the Chinese method is highly effective for measuring footprints of both dust and inked forms. However, as shown by the SEM and LOA results, the reliability of dust footprints was slightly lower, and the difference was more pronounced when it came to rater reliability tests.
A possible explanation is that the borders of dust footprints are generally less distinct, and each rater has their own criteria when identifying the base point for measurement. Consequently, larger intervals may come into being for dust footprints. Nevertheless, the reliability of dust footprint measurements was not substantially lower than that of inked footprints for either the Chinese method or the Reel method.
Comparison between test–retest reliability and rater reliability
In this study, the ICC of a two-way random effect model was employed for rater reliability assessment, which is more appropriate and rigorous when assessing reliability between different raters [25]. Compared with test–retest reliability, the Ball width exhibited a more pronounced declined in rater reliability, suggesting raters tend to apply different criteria when identifying the outer bilateral edges of footprints.
Comparison between the Chinese method and the Reel method
The high reliability of the Reel method is in line with the findings of the previous study [12]. Compared with the Chinese method, the Reel method demonstrated superior reliability. Furthermore, the results confirm that raters following the Reel method protocol can achieve highly reliable measurements. As for the Chinese method, the determination of the fundamental central line relies on identifying the central point of the second toe, and uncertainty may come from re-measurement or measurement by different raters. The IC measure software used in this study can automatically determine the center of the defined circles. Therefore, the middle of the second toe and the outmost of the heel can be determined to be relatively stable.
However, the Reel method did not demonstrate a distinct advantage over the Chinese method. In addition, the Chinese method is simpler to execute in practice. As for the Reel method, footprint photos should be enlarged from the bottom line to create sufficient space for identifying the intersection of the inner and outer tangents—a more complex procedural step. Overall, the Heel width measurement in the Chinese method achieved the highest reliability score, whereas the Arch width yielded the lowest, likeably due to the subjective judgment involved in determining the “widest part of the arch” for this measurement.
In this paper, only the Calc_A1 metric from the Reel method and four primary metrics from the Chinese method were selected for the reliability comparison. Therefore, although the Reel method showed generally superior reliability in this study, it cannot be concluded that it is uniformly more reliable than the Chinese method across all potential metrics. Future studies may therefore include all secondary metrics to enable a comprehensive comparison. In addition, whereas this study used clear and complete dust footprints, further research should evaluate the reliability of both methods on partial or blurred footprints of lower quality.
The absolute reliability metrics presented in this study, including the SEM and LOA, may more directly reflect the variation in measurement during footprint identification in real cases. For researchers investigating the stability and specificity of bare footprints, the reliability test of the measurement methods are particularly valuable. In addition, given its capability in determining measurement reference points, the application of the IC Measure software is recommended in practice for more accurate results.
Conclusion
The Chinese method for bare footprint measurement demonstrates high reliability with a low SEM. Although the Reel method achieved marginally higher reliability scores, its advantage was not statistically significant. Due to its operational simplicity and distinct measurement criteria, the Chinese method can serve as a practical supplement to the Reel method in regions beyond China.
Although inked footprints shows higher measurement reliability than dust footprints, this advantage was not statistically significant. For the Chinese method, the Arch width and Ball width metrics exhibited lower stability compared to other measurements, which are more pronounced when measured by different raters. During bare footprint identification in China, the reliability values presented in this study could serve as a practical reference when re-measured or measured by multiple raters.
Acknowledgements
The authors would like to thank Wei Zhang, Baicheng Chen, and Hengyu Zang for their assistance as raters in the measurement experiments.
Contributor Information
Kai Sun, School of Criminal Science and Technology, Zhejiang Police College, Hangzhou, China.
Yaping Luo, People’s Public Security University of China, Beijing, China.
Authors' contributions
Kai Sun conceived the idea for the study, performed the literature search, collected the data for analysis, and drafted the figures and manuscript. Yaping Luo secured funding for the project and critically reviewed and revised the manuscript. Both authors read and approved the final version of the manuscript.
Compliance with ethical standards
This study was approved by the Zhejiang Police College Ethics Committee (approval number: ZL2023005) and conducted in accordance with its policies. Prior to participation, written informed consent was obtained from all individuals after a full explanation of the study procedures. The consent forms explicitly stated that anonymized data might be used for research publication.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
This study was supported by People’s Public Security University of China [grant number 2023SYL06].
References
- 1. Reel S, Rouse S, Vernon W, et al. Reliability of a two-dimensional footprint measurement approach. Sci Justice. 2010;50:113–118. [DOI] [PubMed] [Google Scholar]
- 2. DiMaggio JA, Vernon DW. Forensic podiatry: principles and methods. 2nd ed. Boca Raton (FL): CRC Press; 2017. [Google Scholar]
- 3. Gunn N. Old and new methods of evaluating footprint impressions by a forensic podiatrist. Br J Podiatr Med Surg. 1991;3:8–11. [Google Scholar]
- 4. Kennedy RB, Pressman IS, Chen S, et al. Statistical analysis of barefoot impressions. J Forensic Sci. 2003;48:55–63. [PubMed] [Google Scholar]
- 5. Kennedy RB, Chen S, Pressman IS, et al. A large-scale statistical analysis of barefoot impressions. J Forensic Sci. 2005;50:1071–1080. [PubMed] [Google Scholar]
- 6. Ministry of Public Security of the People’s Republic of China . GA/T1496–2018: Terminology for Footprint Examination. Beijing (China): China Standards Press; 2018. Chinese.
- 7. Ministry of Public Security of the People’s Republic of China . GA/T1451–2017: Specifications for Classification of Footprint Characteristics in Forensics. Beijing (China): China Standards Press; 2017. Chinese.
- 8. Ministry of Public Security of the People’s Republic of China . GA/T1532–2018: Technical Specifications for Footprints Examination. Beijing (China): China Standards Press; 2018. Chinese.
- 9. Keer J. Soviet crime countermeasures. In: Trace Evidence Science. Beijing (China): China Renmin University Press; 1957. p. 70–96. [Google Scholar]
- 10. Wu XM, Gao YQ. Footprint theory. In: Characteristics of Barefoot Footprint. Beijing (China): Police Officer Education Press; 1996. p. 20–24. Chinese. [Google Scholar]
- 11. Reel S. Development and evaluation of a valid and reliable footprint measurement approach in forensic identification [Dissertation]. York (UK): The University of Leeds, York St John University, 2012. [Google Scholar]
- 12. Nirenberg MS, Ansert E, Reel S. Reliability of a two-dimensional sock-clad footprint linear measurement method. Sci Justice. 2021;61:649–656. [DOI] [PubMed] [Google Scholar]
- 13. President’s Council of Advisors on Science and Technology (PCAST) . Forensic science in criminal courts: ensuring scientific validity of feature-comparison methods. Washington, DC: Executive Office of the President; 2016. [Google Scholar]
- 14. Mukhra R, Krishan K, Kanchan T. Bare footprint metric analysis methods for comparison and identification in forensic examinations: a review of literature. J Forensic Leg Med. 2018;58:101–112. [DOI] [PubMed] [Google Scholar]
- 15. Hu A, Arnold JB, Causby R, et al. The identification and reliability of static and dynamic barefoot impression measurements: a systematic review. Forensic Sci Int. 2018;289:156–164. [DOI] [PubMed] [Google Scholar]
- 16. Nirenberg MS, Krishan K, Kanchan T. A metric study of insole foot impressions in footwear of identical twins. J Forensic Leg Med. 2017;52:116–121. [DOI] [PubMed] [Google Scholar]
- 17. Curran M, Holmes I. A comparison of the length and width of static inked two-dimensional bare footprints found on a hard compared to a soft surface. Sci Justice. 2019;59:448–451. [DOI] [PubMed] [Google Scholar]
- 18. Nirenberg MS, Ansert E, Krishan K, et al. Two-dimensional metric comparisons between dynamic bare footprints and insole foot impressions—forensic implications. Sci Justice. 2020;60:145–150. [DOI] [PubMed] [Google Scholar]
- 19. Nirenberg MS, Ansert E, Krishan K, et al. Two-dimensional linear analysis of dynamic bare footprints: a comparison of measurement techniques. Sci Justice. 2019;59:552–557. [DOI] [PubMed] [Google Scholar]
- 20. Howsam N, Bridgen A. A comparative study of standing fleshed foot and walking and jumping bare footprint measurements. Sci Justice. 2018;58:346–354. [DOI] [PubMed] [Google Scholar]
- 21. Burrow JG. A comparison between manual and digital measurements in exemplar bare footprint analysis. Int J Forens Sci. 2017;2:000116. [Google Scholar]
- 22. Burrow JG. Ghosting of images in barefoot exemplar prints collection: issues for analyses. J Forensic Identif. 2015;65:884–900. [Google Scholar]
- 23. Obe WV, Simmonite N, Reel S, et al. An investigation into the cause of the inner dark areas and outer lighter areas (ghosting) seen in dynamically-created two-dimensional bare footprints. Sci Justice. 2017;57:276–282. [DOI] [PubMed] [Google Scholar]
- 24. McGraw KO, Wong SP. Forming inferences about some intraclass correlation coefficients. Psychol Methods. 1996;1:30–46. [Google Scholar]
- 25. Atkinson G, Nevill AM. Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med. 1998;26:217–238. [DOI] [PubMed] [Google Scholar]
- 26. Bruton A, Conway JH, Holgate ST. Reliability: what is it, and how is it measured? Physiotherapy. 2000;86:94–99. [Google Scholar]
- 27. Kottner J, Audige L, Brorson S, et al. Guidelines for reporting reliability and agreement studies (GRRAS) were proposed. Int J Nurs Stud. 2011;48:661–671. [DOI] [PubMed] [Google Scholar]