際際滷

際際滷Share a Scribd company logo
Forensic Biometric Identification Solutions LLC.
Realizing an attorney can't be expected  to keep up with all the information related to the  forensic disciplines, we decided to do a series of  brief informational presentations on  latent fingerprint images. Fingerprints and Automated Fingerprint Identification Systems  are an accurate and critical part of the judicial system, but they do have shortcomings  which can result in incorrect fingerprint analysis and sometimes missed/bad identifications.  Errors can occur as a result of fingerprint images not being analyzed thoroughly or when a  fingerprint examiner makes a questionable or bad decision. The current adversarial  judicial process is the most effective tool in minimizing the number of incomplete and  inaccurate latent fingerprint errors, but for the system to work as intended it requires  that attorneys have more useable information available on analyzing/processing latent  fingerprints. As explained later in this presentation a missed suspect or victim exclusion  can be as devastating to a case as a bad identification. This presentation is designed to  provide information on the initial evaluation of the latent image and areas in the fingerprint  process that should be weighed when confronted with  a case involving latent images. The  questions posed in this presentation should be as useful to the prosecution as a they are  to a defense attorney, neither attorney wants to have any surprises in court.
The  image does not need to be verifiable to be useable. If a determination of pattern or  even if limited amount of minutiae detail is available the image can have value in excluding  possible donors. Exclusion is the decision by a fingerprint examiner that there are sufficient  features in disagreement to conclude that the two areas of friction ridge impressions did not  originate from the same source. In some situations knowing the suspect or victim did not  leave the latent image can be very significant in a case.  Scientific Working Group on Friction Ridge Analysis Study and Technology (SWGFAST) guidelines: The standard for exclusion is disagreement of friction ridge details, the following conditions that must be satisfied: Determined by a competent examiner. Applied to all comparable anatomical areas. Presence of one *discrepancy is sufficient to exclude. Based on sufficient quantity and quality of the friction ridge details. * Discrepancy- The presence of friction ridge detail in one impression that does not exist in the corresponding area of another impression.
There are no national standard guidelines for the evaluation of latent images. There are  commonly two approaches to the Determination of Value of Latent Images that are  determined by Agency Policy. Approach #1: Only impressions of value for individualization are compared.  Value for  individualization indicates an impression that is deemed to be identifiable. When adopting this approach, impressions lacking value for individualization are not further compared. Approach #2: Impressions of value for individualization or exclusion are compared.  Approach #1 limits the examiners ability to thoroughly process and evaluate the latent case,  this can lead to an exclusion of information critical to a case. The ability to exclude can be a  an integral part of a case.
The analysis and evaluation of the latent image to determine usability can be done at the  crime scene or in the ID Laboratory based on Agency Policy. A determination on the image  value made at the crime scene could result in latent images deemed no value being destroyed or filed  away in a case folder. Evaluating and Analyzing the latent images at the ID laboratory will  normally produce the best results. The ID Laboratory has the capability to apply software  enhancements, utilize fiber optic lighting and magnification technique to assist in analyzing  and evaluating the questionable latent image(s). Again while the images may not be  identifiable they may contain enough information to exclude a victim or suspect. If the  image from the crime scene can exclude the suspect it may be a significant factor in processing  the latent case. It is important that the attorneys involved in the case thoroughly understand the  process used by the agency in; evaluating latent images and whether the image(s) were categorized  as useable or of no value and if policy dictates the no value images be destroyed or archived.
Fingerprint experts can reach different opinions on the same images? The experts analysis is  based on training and experience this can result in a more experienced examiner placing a  value on an image that the inexperienced examiner may determine has no value. The  determination of image value is not a completely objective decision:  1) Value versus no value 2) Individualization versus inconclusive 3) Exclusion versus inconclusive 4) Individualization versus exclusion The criminal justice system relies on the skill of latent print examiners as experts, currently there is  no generally accepted objective, measure to assess the skill of latent print examiners. For an initial  analysis of incoming latent images an agency may utilize its most experienced examiner or it may  utilize the least experienced, as a training exercise for the new examiner. The training and experience  of the examiner can have a dramatic impact on the images an agency categorizes as no-value or  useable.
Documentation of the image analysis and disposition of the latent image(s) again is based on  Agency Policy. Many agencies document the final decision of the examiner in a computerized  record but few go into detail on how they reached the no-value decision. Remember unlike  individualizations the no-value decision is usually made by a single fingerprint examiner so  you are relying on that individuals; training and experience. If the decision on image quality is  made at the crime scene the individual may not be made by an experienced latent examiner  that determines the value of the image and the conditions at the scene will not be conducive  to a thorough accurate analysis.
Level 1  includes the general ridge flow and pattern configuration to include information enabling orientation, core and delta location, and distinction of finger versus palm.  Level 2  detail includes formations, defined as a ridge ending, bifurcation, dot, or combinations thereof. The relationship of Level 2 detail enables individualization.  Level 3  detail includes all dimensional attributes of a ridge, such as ridge path deviation, width, shape, pores, edge contour, incipient ridges, breaks, creases, scars and other permanent details.
SWGFAST Guidelines; an exclusion decision can be based on: Level 1 detail, when sufficient pattern area and orientation indicators are available  absent any significant distortion. If significant distortion is observed, an exclusion  decision can only be reached by considering both Level 1 and Level 2 details. Level 2 detail, an exclusion decision can be based on Level 2 when sufficient  disagreement has been observed. Level 3 details, cannot be the sole factor in exclusion decisions. Level 3 details  have to be considered in conjunction with Level 1 and Level 2 details.
Images in the low quantity of minutiae may still have enough detail to warrant exclusion
While we would be more than willing to assist you in any case work needs, we are hope this  presentation will give you some basic information for any of your future latent fingerprint  case work. The information contained attorneys working for the prosecution or defense. If  you need more detailed information on latent fingerprints please contact us, in  todays electronic world Forensic Biometric Identification Solutions LLC. can assist  attorneys in NY or CA as easily as we can attorneys in Omaha NE.
油
Exclusion  The determination by an examiner that there is sufficient quality and quantity of detail in disagreement to conclude that two areas of friction ridge impressions did not originate from the same source.  No Value(Inconclusive)  The determination by an examiner that there is neither sufficient agreement to individualize, nor sufficient disagreement to exclude.  Individualization  The determination by an examiner that there is sufficient quality and quantity of detail in agreement to conclude that two friction ridge impressions originated from the same source.  Level 1  includes the general ridge flow and pattern configuration to include information enabling orientation, core and delta location, and distinction of finger versus palm.  Level 2  detail includes formations, defined as a ridge ending, bifurcation, dot, or combinations thereof. The relationship of Level 2 detail enables individualization.  Level 3  detail includes all dimensional attributes of a ridge, such as ridge path deviation, width, shape, pores, edge contour, incipient ridges, breaks, creases, scars and other permanent details.
Ad

Recommended

Latent Fingerprint Matching using Descriptor Based Hough Tranform
Latent Fingerprint Matching using Descriptor Based Hough Tranform
vishakhmarari
Detection and rectification of distorted fingerprint
Detection and rectification of distorted fingerprint
Jayakrishnan U
Latent Fingerprint Individualization
Latent Fingerprint Individualization
Rmcauley
As An Attorney What You Should Expect From A Latent Expert
As An Attorney What You Should Expect From A Latent Expert
Rmcauley
Cd For Attorneys
Cd For Attorneys
Rmcauley
Verification Small Sample from one Training Module
Verification Small Sample from one Training Module
Rmcauley
Seminar, Webinars and Training available from Forensic Biometric Identificati...
Seminar, Webinars and Training available from Forensic Biometric Identificati...
Rmcauley
Unidentified Latent Images
Unidentified Latent Images
Rmcauley
Automated Fingerprint Identification Systems
Automated Fingerprint Identification Systems
Rmcauley
Swot Analysis in Methods of Developing Latent Prints
Swot Analysis in Methods of Developing Latent Prints
JENNIEOLANO
Leading Fingerprint Evidence in Trinidad and Tobago
Leading Fingerprint Evidence in Trinidad and Tobago
Brent Winter
Latent fingerprint and vein matching using ridge feature identification
Latent fingerprint and vein matching using ridge feature identification
eSAT Publishing House
ICDP 2011
ICDP 2011
nshort21
basic latent print lecture for criminology
basic latent print lecture for criminology
Ano Nika
Fingerprint, seminar at IASRI, New Delhi
Fingerprint, seminar at IASRI, New Delhi
Nishikant Taksande
Fingerprint recognition presentation
Fingerprint recognition presentation
Vivek Kumar
Adaptive and Faster Approach to Fingerprint Minutiae Extraction and Validation
Adaptive and Faster Approach to Fingerprint Minutiae Extraction and Validation
CSCJournals
Fingerprint recognition
Fingerprint recognition
ranjit banshpal
Fingerprint recognition
Fingerprint recognition
Tufail Ahmed
An Introduction to Investigative Face Recognition and Animetrics ForensicaGPS
An Introduction to Investigative Face Recognition and Animetrics ForensicaGPS
Animetrics
Animetrics Webinar: How Law Enforcement is Closing Cases with Face Biometric ...
Animetrics Webinar: How Law Enforcement is Closing Cases with Face Biometric ...
Animetrics
PERSONAL-IDENTIFICATION.CHAPTER-7.pptx..
PERSONAL-IDENTIFICATION.CHAPTER-7.pptx..
Marco908326
Paulino fengjain latentfp_matching_descriptorbasedhoughtransform_ijcb11
Paulino fengjain latentfp_matching_descriptorbasedhoughtransform_ijcb11
vishakhmarari
Friction Skin Identification
Friction Skin Identification
Jury Rocamora
Fingerprint recognition
Fingerprint recognition
varsha mohite
Fingerprint presentation...
Fingerprint presentation...
Stina14
Iaetsd latent fingerprint recognition and matching
Iaetsd latent fingerprint recognition and matching
Iaetsd Iaetsd
Finger print recognition
Finger print recognition
Karam Munir Butt
From Manual to Auto Searching- FME in the Driver's Seat
From Manual to Auto Searching- FME in the Driver's Seat
Safe Software
Improving Data Integrity: Synchronization between EAM and ArcGIS Utility Netw...
Improving Data Integrity: Synchronization between EAM and ArcGIS Utility Netw...
Safe Software

More Related Content

Similar to No Nalue Latent Images (20)

Automated Fingerprint Identification Systems
Automated Fingerprint Identification Systems
Rmcauley
Swot Analysis in Methods of Developing Latent Prints
Swot Analysis in Methods of Developing Latent Prints
JENNIEOLANO
Leading Fingerprint Evidence in Trinidad and Tobago
Leading Fingerprint Evidence in Trinidad and Tobago
Brent Winter
Latent fingerprint and vein matching using ridge feature identification
Latent fingerprint and vein matching using ridge feature identification
eSAT Publishing House
ICDP 2011
ICDP 2011
nshort21
basic latent print lecture for criminology
basic latent print lecture for criminology
Ano Nika
Fingerprint, seminar at IASRI, New Delhi
Fingerprint, seminar at IASRI, New Delhi
Nishikant Taksande
Fingerprint recognition presentation
Fingerprint recognition presentation
Vivek Kumar
Adaptive and Faster Approach to Fingerprint Minutiae Extraction and Validation
Adaptive and Faster Approach to Fingerprint Minutiae Extraction and Validation
CSCJournals
Fingerprint recognition
Fingerprint recognition
ranjit banshpal
Fingerprint recognition
Fingerprint recognition
Tufail Ahmed
An Introduction to Investigative Face Recognition and Animetrics ForensicaGPS
An Introduction to Investigative Face Recognition and Animetrics ForensicaGPS
Animetrics
Animetrics Webinar: How Law Enforcement is Closing Cases with Face Biometric ...
Animetrics Webinar: How Law Enforcement is Closing Cases with Face Biometric ...
Animetrics
PERSONAL-IDENTIFICATION.CHAPTER-7.pptx..
PERSONAL-IDENTIFICATION.CHAPTER-7.pptx..
Marco908326
Paulino fengjain latentfp_matching_descriptorbasedhoughtransform_ijcb11
Paulino fengjain latentfp_matching_descriptorbasedhoughtransform_ijcb11
vishakhmarari
Friction Skin Identification
Friction Skin Identification
Jury Rocamora
Fingerprint recognition
Fingerprint recognition
varsha mohite
Fingerprint presentation...
Fingerprint presentation...
Stina14
Iaetsd latent fingerprint recognition and matching
Iaetsd latent fingerprint recognition and matching
Iaetsd Iaetsd
Finger print recognition
Finger print recognition
Karam Munir Butt
Automated Fingerprint Identification Systems
Automated Fingerprint Identification Systems
Rmcauley
Swot Analysis in Methods of Developing Latent Prints
Swot Analysis in Methods of Developing Latent Prints
JENNIEOLANO
Leading Fingerprint Evidence in Trinidad and Tobago
Leading Fingerprint Evidence in Trinidad and Tobago
Brent Winter
Latent fingerprint and vein matching using ridge feature identification
Latent fingerprint and vein matching using ridge feature identification
eSAT Publishing House
ICDP 2011
ICDP 2011
nshort21
basic latent print lecture for criminology
basic latent print lecture for criminology
Ano Nika
Fingerprint, seminar at IASRI, New Delhi
Fingerprint, seminar at IASRI, New Delhi
Nishikant Taksande
Fingerprint recognition presentation
Fingerprint recognition presentation
Vivek Kumar
Adaptive and Faster Approach to Fingerprint Minutiae Extraction and Validation
Adaptive and Faster Approach to Fingerprint Minutiae Extraction and Validation
CSCJournals
Fingerprint recognition
Fingerprint recognition
ranjit banshpal
Fingerprint recognition
Fingerprint recognition
Tufail Ahmed
An Introduction to Investigative Face Recognition and Animetrics ForensicaGPS
An Introduction to Investigative Face Recognition and Animetrics ForensicaGPS
Animetrics
Animetrics Webinar: How Law Enforcement is Closing Cases with Face Biometric ...
Animetrics Webinar: How Law Enforcement is Closing Cases with Face Biometric ...
Animetrics
PERSONAL-IDENTIFICATION.CHAPTER-7.pptx..
PERSONAL-IDENTIFICATION.CHAPTER-7.pptx..
Marco908326
Paulino fengjain latentfp_matching_descriptorbasedhoughtransform_ijcb11
Paulino fengjain latentfp_matching_descriptorbasedhoughtransform_ijcb11
vishakhmarari
Friction Skin Identification
Friction Skin Identification
Jury Rocamora
Fingerprint recognition
Fingerprint recognition
varsha mohite
Fingerprint presentation...
Fingerprint presentation...
Stina14
Iaetsd latent fingerprint recognition and matching
Iaetsd latent fingerprint recognition and matching
Iaetsd Iaetsd
Finger print recognition
Finger print recognition
Karam Munir Butt

Recently uploaded (20)

From Manual to Auto Searching- FME in the Driver's Seat
From Manual to Auto Searching- FME in the Driver's Seat
Safe Software
Improving Data Integrity: Synchronization between EAM and ArcGIS Utility Netw...
Improving Data Integrity: Synchronization between EAM and ArcGIS Utility Netw...
Safe Software
GenAI Opportunities and Challenges - Where 370 Enterprises Are Focusing Now.pdf
GenAI Opportunities and Challenges - Where 370 Enterprises Are Focusing Now.pdf
Priyanka Aash
Securing AI - There Is No Try, Only Do!.pdf
Securing AI - There Is No Try, Only Do!.pdf
Priyanka Aash
Enhance GitHub Copilot using MCP - Enterprise version.pdf
Enhance GitHub Copilot using MCP - Enterprise version.pdf
Nilesh Gule
MuleSoft for AgentForce : Topic Center and API Catalog
MuleSoft for AgentForce : Topic Center and API Catalog
shyamraj55
The Future of Technology: 2025-2125 by Saikat Basu.pdf
The Future of Technology: 2025-2125 by Saikat Basu.pdf
Saikat Basu
Smarter Aviation Data Management: Lessons from Swedavia Airports and Sweco
Smarter Aviation Data Management: Lessons from Swedavia Airports and Sweco
Safe Software
2025_06_18 - OpenMetadata Community Meeting.pdf
2025_06_18 - OpenMetadata Community Meeting.pdf
OpenMetadata
Connecting Data and Intelligence: The Role of FME in Machine Learning
Connecting Data and Intelligence: The Role of FME in Machine Learning
Safe Software
The Future of AI Agent Development Trends to Watch.pptx
The Future of AI Agent Development Trends to Watch.pptx
Lisa ward
Creating Inclusive Digital Learning with AI: A Smarter, Fairer Future
Creating Inclusive Digital Learning with AI: A Smarter, Fairer Future
Impelsys Inc.
10 Key Challenges for AI within the EU Data Protection Framework.pdf
10 Key Challenges for AI within the EU Data Protection Framework.pdf
Priyanka Aash
Techniques for Automatic Device Identification and Network Assignment.pdf
Techniques for Automatic Device Identification and Network Assignment.pdf
Priyanka Aash
FIDO Seminar: Evolving Landscape of Post-Quantum Cryptography.pptx
FIDO Seminar: Evolving Landscape of Post-Quantum Cryptography.pptx
FIDO Alliance
Python Conference Singapore - 19 Jun 2025
Python Conference Singapore - 19 Jun 2025
ninefyi
FIDO Seminar: New Data: Passkey Adoption in the Workforce.pptx
FIDO Seminar: New Data: Passkey Adoption in the Workforce.pptx
FIDO Alliance
War_And_Cyber_3_Years_Of_Struggle_And_Lessons_For_Global_Security.pdf
War_And_Cyber_3_Years_Of_Struggle_And_Lessons_For_Global_Security.pdf
biswajitbanerjee38
ReSTIR [DI]: Spatiotemporal reservoir resampling for real-time ray tracing ...
ReSTIR [DI]: Spatiotemporal reservoir resampling for real-time ray tracing ...
revolcs10
FIDO Alliance Seminar State of Passkeys.pptx
FIDO Alliance Seminar State of Passkeys.pptx
FIDO Alliance
From Manual to Auto Searching- FME in the Driver's Seat
From Manual to Auto Searching- FME in the Driver's Seat
Safe Software
Improving Data Integrity: Synchronization between EAM and ArcGIS Utility Netw...
Improving Data Integrity: Synchronization between EAM and ArcGIS Utility Netw...
Safe Software
GenAI Opportunities and Challenges - Where 370 Enterprises Are Focusing Now.pdf
GenAI Opportunities and Challenges - Where 370 Enterprises Are Focusing Now.pdf
Priyanka Aash
Securing AI - There Is No Try, Only Do!.pdf
Securing AI - There Is No Try, Only Do!.pdf
Priyanka Aash
Enhance GitHub Copilot using MCP - Enterprise version.pdf
Enhance GitHub Copilot using MCP - Enterprise version.pdf
Nilesh Gule
MuleSoft for AgentForce : Topic Center and API Catalog
MuleSoft for AgentForce : Topic Center and API Catalog
shyamraj55
The Future of Technology: 2025-2125 by Saikat Basu.pdf
The Future of Technology: 2025-2125 by Saikat Basu.pdf
Saikat Basu
Smarter Aviation Data Management: Lessons from Swedavia Airports and Sweco
Smarter Aviation Data Management: Lessons from Swedavia Airports and Sweco
Safe Software
2025_06_18 - OpenMetadata Community Meeting.pdf
2025_06_18 - OpenMetadata Community Meeting.pdf
OpenMetadata
Connecting Data and Intelligence: The Role of FME in Machine Learning
Connecting Data and Intelligence: The Role of FME in Machine Learning
Safe Software
The Future of AI Agent Development Trends to Watch.pptx
The Future of AI Agent Development Trends to Watch.pptx
Lisa ward
Creating Inclusive Digital Learning with AI: A Smarter, Fairer Future
Creating Inclusive Digital Learning with AI: A Smarter, Fairer Future
Impelsys Inc.
10 Key Challenges for AI within the EU Data Protection Framework.pdf
10 Key Challenges for AI within the EU Data Protection Framework.pdf
Priyanka Aash
Techniques for Automatic Device Identification and Network Assignment.pdf
Techniques for Automatic Device Identification and Network Assignment.pdf
Priyanka Aash
FIDO Seminar: Evolving Landscape of Post-Quantum Cryptography.pptx
FIDO Seminar: Evolving Landscape of Post-Quantum Cryptography.pptx
FIDO Alliance
Python Conference Singapore - 19 Jun 2025
Python Conference Singapore - 19 Jun 2025
ninefyi
FIDO Seminar: New Data: Passkey Adoption in the Workforce.pptx
FIDO Seminar: New Data: Passkey Adoption in the Workforce.pptx
FIDO Alliance
War_And_Cyber_3_Years_Of_Struggle_And_Lessons_For_Global_Security.pdf
War_And_Cyber_3_Years_Of_Struggle_And_Lessons_For_Global_Security.pdf
biswajitbanerjee38
ReSTIR [DI]: Spatiotemporal reservoir resampling for real-time ray tracing ...
ReSTIR [DI]: Spatiotemporal reservoir resampling for real-time ray tracing ...
revolcs10
FIDO Alliance Seminar State of Passkeys.pptx
FIDO Alliance Seminar State of Passkeys.pptx
FIDO Alliance
Ad

No Nalue Latent Images

  • 2. Realizing an attorney can't be expected to keep up with all the information related to the forensic disciplines, we decided to do a series of brief informational presentations on latent fingerprint images. Fingerprints and Automated Fingerprint Identification Systems are an accurate and critical part of the judicial system, but they do have shortcomings which can result in incorrect fingerprint analysis and sometimes missed/bad identifications. Errors can occur as a result of fingerprint images not being analyzed thoroughly or when a fingerprint examiner makes a questionable or bad decision. The current adversarial judicial process is the most effective tool in minimizing the number of incomplete and inaccurate latent fingerprint errors, but for the system to work as intended it requires that attorneys have more useable information available on analyzing/processing latent fingerprints. As explained later in this presentation a missed suspect or victim exclusion can be as devastating to a case as a bad identification. This presentation is designed to provide information on the initial evaluation of the latent image and areas in the fingerprint process that should be weighed when confronted with a case involving latent images. The questions posed in this presentation should be as useful to the prosecution as a they are to a defense attorney, neither attorney wants to have any surprises in court.
  • 3. The image does not need to be verifiable to be useable. If a determination of pattern or even if limited amount of minutiae detail is available the image can have value in excluding possible donors. Exclusion is the decision by a fingerprint examiner that there are sufficient features in disagreement to conclude that the two areas of friction ridge impressions did not originate from the same source. In some situations knowing the suspect or victim did not leave the latent image can be very significant in a case. Scientific Working Group on Friction Ridge Analysis Study and Technology (SWGFAST) guidelines: The standard for exclusion is disagreement of friction ridge details, the following conditions that must be satisfied: Determined by a competent examiner. Applied to all comparable anatomical areas. Presence of one *discrepancy is sufficient to exclude. Based on sufficient quantity and quality of the friction ridge details. * Discrepancy- The presence of friction ridge detail in one impression that does not exist in the corresponding area of another impression.
  • 4. There are no national standard guidelines for the evaluation of latent images. There are commonly two approaches to the Determination of Value of Latent Images that are determined by Agency Policy. Approach #1: Only impressions of value for individualization are compared. Value for individualization indicates an impression that is deemed to be identifiable. When adopting this approach, impressions lacking value for individualization are not further compared. Approach #2: Impressions of value for individualization or exclusion are compared. Approach #1 limits the examiners ability to thoroughly process and evaluate the latent case, this can lead to an exclusion of information critical to a case. The ability to exclude can be a an integral part of a case.
  • 5. The analysis and evaluation of the latent image to determine usability can be done at the crime scene or in the ID Laboratory based on Agency Policy. A determination on the image value made at the crime scene could result in latent images deemed no value being destroyed or filed away in a case folder. Evaluating and Analyzing the latent images at the ID laboratory will normally produce the best results. The ID Laboratory has the capability to apply software enhancements, utilize fiber optic lighting and magnification technique to assist in analyzing and evaluating the questionable latent image(s). Again while the images may not be identifiable they may contain enough information to exclude a victim or suspect. If the image from the crime scene can exclude the suspect it may be a significant factor in processing the latent case. It is important that the attorneys involved in the case thoroughly understand the process used by the agency in; evaluating latent images and whether the image(s) were categorized as useable or of no value and if policy dictates the no value images be destroyed or archived.
  • 6. Fingerprint experts can reach different opinions on the same images? The experts analysis is based on training and experience this can result in a more experienced examiner placing a value on an image that the inexperienced examiner may determine has no value. The determination of image value is not a completely objective decision: 1) Value versus no value 2) Individualization versus inconclusive 3) Exclusion versus inconclusive 4) Individualization versus exclusion The criminal justice system relies on the skill of latent print examiners as experts, currently there is no generally accepted objective, measure to assess the skill of latent print examiners. For an initial analysis of incoming latent images an agency may utilize its most experienced examiner or it may utilize the least experienced, as a training exercise for the new examiner. The training and experience of the examiner can have a dramatic impact on the images an agency categorizes as no-value or useable.
  • 7. Documentation of the image analysis and disposition of the latent image(s) again is based on Agency Policy. Many agencies document the final decision of the examiner in a computerized record but few go into detail on how they reached the no-value decision. Remember unlike individualizations the no-value decision is usually made by a single fingerprint examiner so you are relying on that individuals; training and experience. If the decision on image quality is made at the crime scene the individual may not be made by an experienced latent examiner that determines the value of the image and the conditions at the scene will not be conducive to a thorough accurate analysis.
  • 8. Level 1 includes the general ridge flow and pattern configuration to include information enabling orientation, core and delta location, and distinction of finger versus palm. Level 2 detail includes formations, defined as a ridge ending, bifurcation, dot, or combinations thereof. The relationship of Level 2 detail enables individualization. Level 3 detail includes all dimensional attributes of a ridge, such as ridge path deviation, width, shape, pores, edge contour, incipient ridges, breaks, creases, scars and other permanent details.
  • 9. SWGFAST Guidelines; an exclusion decision can be based on: Level 1 detail, when sufficient pattern area and orientation indicators are available absent any significant distortion. If significant distortion is observed, an exclusion decision can only be reached by considering both Level 1 and Level 2 details. Level 2 detail, an exclusion decision can be based on Level 2 when sufficient disagreement has been observed. Level 3 details, cannot be the sole factor in exclusion decisions. Level 3 details have to be considered in conjunction with Level 1 and Level 2 details.
  • 10. Images in the low quantity of minutiae may still have enough detail to warrant exclusion
  • 11. While we would be more than willing to assist you in any case work needs, we are hope this presentation will give you some basic information for any of your future latent fingerprint case work. The information contained attorneys working for the prosecution or defense. If you need more detailed information on latent fingerprints please contact us, in todays electronic world Forensic Biometric Identification Solutions LLC. can assist attorneys in NY or CA as easily as we can attorneys in Omaha NE.
  • 12.
  • 13. Exclusion The determination by an examiner that there is sufficient quality and quantity of detail in disagreement to conclude that two areas of friction ridge impressions did not originate from the same source. No Value(Inconclusive) The determination by an examiner that there is neither sufficient agreement to individualize, nor sufficient disagreement to exclude. Individualization The determination by an examiner that there is sufficient quality and quantity of detail in agreement to conclude that two friction ridge impressions originated from the same source. Level 1 includes the general ridge flow and pattern configuration to include information enabling orientation, core and delta location, and distinction of finger versus palm. Level 2 detail includes formations, defined as a ridge ending, bifurcation, dot, or combinations thereof. The relationship of Level 2 detail enables individualization. Level 3 detail includes all dimensional attributes of a ridge, such as ridge path deviation, width, shape, pores, edge contour, incipient ridges, breaks, creases, scars and other permanent details.

Editor's Notes

  • #13: While you are doing the exercises I will pass out by business card.