Comparative analysis of machine learning algorithms for Lip print based person identification

Sandhya, Sankaran; Fernandes, Roshan; Sapna, S.; Rodrigues, Anisha P

doi:10.1007/s12065-020-00561-y

Cited by 11 publications

(6 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These models incorporate various lip print features, such as patterns and grooves, leading to more comprehensive and accurate analysis compared to human examination alone [ 20 ]. Notably, comparative analysis using AI algorithms facilitates efficient matching against extensive databases, aiding forensic investigations and identification procedures [ 21 ].…”

Section: Discussionmentioning

confidence: 99%

Evaluating Digital Photography for Lip Print Recording: Compatibility With Traditional Classification Systems

George Pallivathukal,

Kumar,

Joy Idiculla

et al. 2024

Cureus

View full text Add to dashboard Cite

Background: Direct digital photography for lip print recording is a recent trend, and there is a dearth of systematic research on the analysis of the recorded prints with existing clip print classification systems.Aims and objectives: This study aims to compare the accuracy of the digital photographic method in lip print recording, comparing it with traditional methods, and assessing the suitability of commonly used lip print classification for analyzing lip prints recorded by photographic method.Methodology: A total of 72 participants aged between 20 and 26 were included. The lip print recording process involved photographing the lips without and with lipstick, followed by recording the lip print with cellophane tape on bond paper. The prints collected using the different methods were analyzed and compared for agreement, and the data were analyzed statistically.Results: Inter-observer reliability was high for all methods (>0.800). The distribution of lip print patterns differed across the methods, suggesting a potential influence of the recording technique. The agreement between the conventional method and both digital methods was moderate (kappa=0.449-0.517). The agreement between digital methods with and without enhancement was also moderate (kappa=0.718). Notably, digital photographs with enhancement tend to have a higher positive agreement for several lip print types.Conclusion: Digital photography is a potential method for lip print recording. However, this study highlights the need for the calibration of lip print classification systems for digitally recorded lip prints. Further research is needed to refine the use of digital photography in forensic lip print analysis and to explore its integration with artificial intelligence for biometric identification.

show abstract

Section: Discussionmentioning

confidence: 99%

Evaluating Digital Photography for Lip Print Recording: Compatibility With Traditional Classification Systems

George Pallivathukal,

Kumar,

Joy Idiculla

et al. 2024

Cureus

View full text Add to dashboard Cite

show abstract

“…They reported an accuracy above 94% on a database of 1001 lip prints taken from 143 individuals (7 lip prints per person) with a technique allowing to detect low quality areas in the prints. Sandhya et al [ 488 ] obtained an accuracy of 97% on a smaller dataset (300 individuals) using an ensemble classifiers informed by three classifiers (KNN, SVM, ANN).…”

Section: Other Body Marks Case Reports and Miscellaneousmentioning

confidence: 99%

Interpol review of fingermarks and other body impressions (2019 – 2022)

Bécue

Champod

2023

Forensic Science International: Synergy

View full text Add to dashboard Cite

“…The authors used 100 lip prints, and the accuracy was 86.7%. Sandhya et al [26] used lip prints for personal identification. LBP was used to obtain texture features from the extracted segmented upper and lower lip.…”

Section: (Lbp)mentioning

confidence: 99%

A Comparison of Methods for Extracting Lip Print Features

عنائي,

Mersal,

مصطفى

2023

النشرة المعلوماتية في الحاسبات والمعلومات

View full text Add to dashboard Cite

Lip prints and other biometric methods are increasingly used to identify people at crime scenes. Since the early 2000s, researchers have shown increased interest in using these methods. A lip print is biometric, like fingerprints, iris, or another feature in the body. Lip prints are unique to everyone, so they do not change over time. The images are classified through various methods to extract the features from the image based on color, shape, and texture features. A texture feature was the most critical extraction feature. This study examines the challenges and accuracy of previous lip print biometric methods and the various available techniques. In this survey, most researchers used feature extraction based on texture techniques like Hough Transform (HT), Dynamic Time Wrapping (DTW), statistical analysis, edge detection operators, and Local Binary Patterns (LBP). Our literature review covered the past ten years. Hough Transform was the most accurate at 98.49%, followed by Local Binary Patterns and Triangle feature set-based fuzzy at 97%.

show abstract

Comparative analysis of machine learning algorithms for Lip print based person identification

Cited by 11 publications

References 27 publications

Evaluating Digital Photography for Lip Print Recording: Compatibility With Traditional Classification Systems

Evaluating Digital Photography for Lip Print Recording: Compatibility With Traditional Classification Systems

Interpol review of fingermarks and other body impressions (2019 – 2022)

A Comparison of Methods for Extracting Lip Print Features

Contact Info

Product

Resources

About