Chapter 7 Forensics

Biometrics

A form of access control that accurately and efficiently identifies humans

• uses an individual’s biological and behavioral traits to grant access

Functions of biometrics

Biometric matching/verification

Biometric Identification

Biometric matching/verification

identifies someone out of a crowd by scanning select biometric characteristics into a database

• ex. law enforcement looking for a person of interest

Biometric identification

Access control systems create and store databases of biological traits and compare them to the individual attempting to gain entry into a device or facility

• ex. facial or retina scan- system decides who gains entry to facility

Types of forensic biometrics

• physiological biometrics

• behavioral biometrics

• iris biometrics

• retina biometrics

Physiological biometrics

focuses on identifying humans through their unique physiological traits

• fingerprints

• hand scans

• iris scans

• retina scans

• facial scans

Behavioral biometrics

focuses on identifying humans through unique behavioral or mental patterns they exhibit- how a person acts (less stable and not used often)

• handwriting- analyzing strokes, curves, pressure of wriitng

• voice- measures pitch/dynamics of voice

• keystroke recognition- -measures speed of pressing keys (good for hackers)

• Gait recognition- cycle of walking, step length, posture etc.

What is the workflow for collecting biometric data?

• enrollment

• Extraction

• Template Module

Enrollment

Captures a persons biometric data and stores it in a database for later use

Extraction

Finds patterns in the traits extracted by the sensor by using mathematical equations

• One of the MOST COMPLEX and important components

Template module

saves all of the raw data produced by the feature extraction and putting it into a simple and easy to read format for the system

• compares a user to the rest of the data to determine match

Iris Biometrics

Goal is to create a unique template of the iris (IrisCode) to be compared against a database in the NGI system

• Visible pattern at the front of the eye

Iris

Colored part of the eye around the pupil

• unique to every individual

• Intricate muscle patterns and iris constriction is unique to every person

• iris goes unchanged for the life of user

Retina biometrics

Requires user to be within a fraction of an inch of a high power camear lens that shines a bright light into the eye to see

• blood vessels at the back of the eye

Advantages of iris biometrics?

• less invasive

• iris is individual to everyone

• less likely to be affected by disease

• better technology

Facial recognition

Analyze a variety of facial features

a clear image must eb acquired

system normalizes the image

template is stored

various techniques developed, appearance based and model based

Advantages of facial recognition

• not intrusive like iris biometrics

• beneficial for day to day access/use

• hygienic- no touching required

Disadvantages of facial recognition

• many features compromise the face

• complex models and graphs needed to store all the data used

• can be tricked by a 2D picture

3 Most widely accepted facial recognition techniques

• principle component analysis

• linear discriminant analysis

• elastic bunch graph matching

Principle Component Analysis

Uses thousands of stored images of faces and places them over the face in question. Precise algorithms analyze the face during overlap

• appearance-based

• disadvantage: full frontal face required

Linear Discriminant Analysis

face is placed on a vector, lines from one feature to another are analyzed. Relationship between the lines is analyzed and variations are recorded

• advantage: fast, can accamodate different lighting and facial expressions

• disadvantage: full face image required, requires a large amount of data

• appearance based

Elastic Bunch Graph Matching

sequence of graphs used to map the nonlinear relationships of features on the face. Identifies landmarks like edges of lips, tips of nose etc. Characteristics are assigned then compared.

• model-based

• advanatge: allows for partial and off center pictures to be analysed, maps human face with extreme accuracy

Next -Generation Identification System (NGI)

Provides a broader selection of resources and information for law enforcement agencies across the country

• integrating biometric technology into identification system

• 
  

Interstate Photo System (IPS)

A database containing 30M front-facing mugshot photos of individuals. ALL face photos MUST INCLUDE all ten fingerprints

• maintained by FBI

• 
  

Facial Analysis, Comparison, and Evaluation (FACE) Services

Provides face recognition capabilities to support active FBI investigations

• Contains civil photos from state/federal government databases

National Palm Print System (NPPS)

Contains database of millions of palmprints to be used in criminal investigations

Repository of Individuals of Special Concern (RISC)

Allows officers ti fingerprint individuals they encounter and quickly detect whether they are a threat

• 10 seconds to search and compare against 2.9M prints

• completely mobile

What are the two main functions of biometrics?

• Biometric matching/verification

• Biometric identification

What are the 3 components of biometrics

• enrollment

• extraction

• template modules

Which category of biometrics is more concerned with the way a human performs an action, rather than focusing on a specific physiological trait?

• Which is most common?

Behavioral biometrics- most common is handwriitng analysis

What is one of the main problems with retina biometrics?

Viable images can be affected by disease

What are the three most widely accepted techniques to create a viable reconstruction and analysis of a human face?

• PCA (Principle component analysis)

• LDA (Linear discriminant analysis)

• EBGM (Elastic Bunch Graph Matching)

What is RISC? Why is it beneficial for law enforcement officers?

Mobile database of individuals of special concern

• allows for officers ti quickly identify threats/potential dangers