MRI-based knee image for personal identification

Shamir, Lior

doi:10.1504/ijbm.2013.052943

Cited by 9 publications

(5 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…where Tc is the training set of class c, t is a feature vector from Tc, |x| is the length of the feature vector x, xf is the value of image feature f, Wf is the Fisher discriminant score of feature f, |Tc| is the number of training samples of class c, d(x,c) is the computed distance from a given sample x to class c, and p is the exponent, which is set to −5 as thoroughly discussed in (Orlov et al, 2008). The comprehensive set of image content descriptors and the selection of the informative image features allow the application of the algorithm to a broad variety of complex image data (Shamir et al, 2008a;Shamir et al, 2012;Shamir, 2012;Shamir & Tarakhovsky, 2012;Shamir, Holincheck & Wallin, 2013;Shamir, 2013). Source code for the method is publicly available .…”

Section: Galaxy Morphology Analysis Methodsmentioning

confidence: 99%

Combining Human and Machine Learning for Morphological Analysis of Galaxy Images

Kuminski¹,

George²,

Wallin³

et al. 2014

Publications of the Astronomical Society of the Pacific

Self Cite

View full text Add to dashboard Cite

ABSTRACT. The increasing importance of digital sky surveys collecting many millions of galaxy images has reinforced the need for robust methods that can perform morphological analysis of large galaxy image databases. Citizen science initiatives such as Galaxy Zoo showed that large data sets of galaxy images can be analyzed effectively by nonscientist volunteers, but since databases generated by robotic telescopes grow much faster than the processing power of any group of citizen scientists, it is clear that computer analysis is required. Here, we propose to use citizen science data for training machine learning systems, and show experimental results demonstrating that machine learning systems can be trained with citizen science data. Our findings show that the performance of machine learning depends on the quality of the data, which can be improved by using samples that have a high degree of agreement between the citizen scientists. The source code of the method is publicly available.

show abstract

Section: Galaxy Morphology Analysis Methodsmentioning

confidence: 99%

Combining Human and Machine Learning for Morphological Analysis of Galaxy Images

Kuminski¹,

George²,

Wallin³

et al. 2014

Publications of the Astronomical Society of the Pacific

Self Cite

View full text Add to dashboard Cite

show abstract

“…The International Society of Forensic Radiology and Imaging (ISFRI) has published positional statements regarding the use of conventional radiographs and PMCT for forensic identification [22,23]. However, the potential of MR for comparative identification remains underinvestigated and literature on this topic is scarce [10,13,14]. Overall, the comparison between MR and CT (regardless of whether it is antemortem or postmortem) is more challenging than intramodal comparison (MR to MR or CT to CT) due to the inherent differences between CT, which provides high resolution and edge detail of bone, and MR, which provides high contrast for soft tissue [12].…”

Section: Discussionmentioning

confidence: 99%

“…Moreover, CT allows for comparisons with radiographs [11]. MR imaging of osseous tissues has less resolution and edge detail than radiographs or CT [12] and has thus only rarely been deemed an appropriate tool for radiologic identification [13,14]. Documentation of radiologic identification for personal ID of living individuals is rare in the literature [15], as ID for living individuals typically relies on ID-documents and fingerprints (https://www.fedpol.admin.ch/fedpol/de/home/ sicherheit/personenidentifikation.html).…”

Section: Introductionmentioning

confidence: 99%

Antemortem identification by fusion of MR and CT of the paranasal sinuses

Heimer

Gascho

Gentile

et al. 2017

Forensic Sci Med Pathol

View full text Add to dashboard Cite

Radiologic forensic identification is usually performed by comparing antemortem and postmortem radiographs. While computed tomography (CT) has become a valuable addition to radiologic identification, magnetic resonance (MR) imaging has only rarely been used for this purpose. In our case, identification was accomplished using fused MR- and CT images in a survivor of a gunshot injury to the head. This case supports and highlights the possibility to perform intermodality radiologic identification comparing preexisting MR imaging to subsequently aquired CT data in living (or deceased) humans as long as manual modifications of windowing, color and contrast enable differentiation of the two modalities in the fused image.

show abstract

“…Extending their research from identification using knee Xrays, Shamir et al [69] proposed later their work on using MRI-based knee images for person identification. The authors showed that the accuracy for person identification using knee MRI scans is significantly higher than random.…”

Section: D Magnet Resonance Imagingmentioning

confidence: 99%

Biometric Recognition in 3D Medical Images: A Survey

Fu,

Damer

2023

IEEE Access

View full text Add to dashboard Cite

This work explores the potential applications of biometric recognition in 3D medical imaging data. We investigate various 3D imaging techniques commonly used in the medical domain, including 3D ultrasound imaging, magnetic resonance imaging (MRI), computer tomography (CT) scans, and 3D nearinfrared (NIR) imaging. For each technique, we provide an overview of its working principle and discuss the advantages of integrating biometrics into 3D medical imaging data. Major advantage of using biometrics in this context is motivated by the research that using biometrics could not only increase data security but, more importantly, decrease the mix-up errors in patient's medical data and thus improving patient safety and patient care. Our analysis uncovers certain weaknesses in current algorithms and limitations in existing research. Possible reasons include insufficient data availability, the under-utilization of deep-learning-based approaches to enhance accuracy and performance, and the absence of standardized benchmarking databases to support research. Our survey frames existing works and lead to practical recommendations and motivates efforts to improve the current state of research. Beyond exploring the utilization of biometrics in 3D medical imaging data, our study touches on further potential interactions between them, such as extracting health information from biometric captures. This work is thus the first work to survey and present an overview on works proposing the use of medical images for biometric recognition.

show abstract

MRI-based knee image for personal identification

Cited by 9 publications

References 28 publications

Combining Human and Machine Learning for Morphological Analysis of Galaxy Images

Combining Human and Machine Learning for Morphological Analysis of Galaxy Images

Antemortem identification by fusion of MR and CT of the paranasal sinuses

Biometric Recognition in 3D Medical Images: A Survey

Contact Info

Product

Resources

About