Introduction
Most commonly used biometric trait by humans and is standard practice to use face in passports, IDs, driver license.
Used for person recognition and revealing attributes such as age, ethnicity and emotional state
Important biometric identifier in law enforcement and Human Computer Interaction (HCI)
Non intrusive
What is Facial Recognition?
Facial Recognition Systems are computer-based security systems that are able to automatically detect and identify or verify a person from a digital image or a video frame from a video source
The most common way of doing this is to compare selected facial features from an image and a facial database
Pre-Processing
Can be used in a controlled and uncontrolled environments
In a controlled environment, frontal and profile photographs of human faces are taken, complete with a uniform background and identical poses among participants (Eg. Passport Photo)
Canonical Face Image
Face images are commonly called mug shots. Each mug shot can be manually or automatically cropped to extract a normalized subpart called a canonical face image
Facial Recognition Issues
These relate to an uncontrolled environment:
- More than one or may faces appear
- Lighting conditions may vary
- Facial expressions vary
- Faces appear at different scales, positions and orientations
- Facial hair, make up and headware
- Faces can be occluded
- Aging
Facial Features
Can be organized into three categories:
Level 1
- Details consist of gross facial characteristics that are easily observable.
- These include general geometry of the face and global skin color.
- Such features can be used to quickly discriminate between genders and races
Level 2
- Consist of localized face information such as the structure of face components (Eg. eye)
- These features are essential for accurate recognition and require a higher resolution face image
- The characteristics of local regions of the face can be represented using geometric or texture descriptors
Level 3
- Details consist of unstructured, micro level features on the face, such as scars, freckles, skin discoloration and moles
- One challenging problem where level 3 details may be critical is the discrimination of identical twins
Face Recognition Technology
Face Recognition Technology Involves:
- Analyzing Facial Characteristics
- Storing Features in a Database
- Using them to identify users
Facial Scan Process Flow:
- Sample Capture – Senors
- Feature Extraction – Creation of template
- Template Comparison
- Verification – 1 to 1
- Identification – 1 to many
- Matching
How does it work?
- First step is to recognize a human face and extract it from the rest of the scene
- The system measures nodal points on the face (distance between the eyes, the shape of the cheekbones)
- Nodal points are then compared to the nodal points computed from a database of pictures in order to find a match
Image Acquisition
Image Acquisition/ Sensor – Takes observation develops biometric signature. Eg, Camera
- Image acquired from a sensor can be categorized based on spectral band [visible, infrared, thermal] and nature of the image [2D, 3D, video]
- Automated recognition requires the face data to be in a machine readable format
Image Acquisition – 2D
Predominant source of face images
Large number of sensors and techniques have been developed for acquiring and processing 2D face images
May occlude facial features. this is called self-occlusion
Recognition Algorithms
Can be divided into two main approaches:
- Geometric – Which looks at distinguishing features
- Photo-metric – Statistical approach that distills an image into values and compares the values with the templates to eliminate variances
Popular recognition algorithms:
- Eigenfaces
- Linear Discriminate Analysis
- Hidden Markov Model
Face Recognition Feature Extraction
- Locate eyes, ears, nose and mouth
- Calculate distances and ratios to common reference points
Image Pre-Processing in Face Recognition
- Detect Face
- Rotate Face
- Pose Correction
- Light/Shade Correction
- Scaling
Leave a Reply