Technologies
for Face Recognition in Access Control
In the
1960s, Woodrow Wilson Bledsoe developed a system that could classify photos of
faces called facial recognition. Identifying human faces in digital images
has variety of applications, from biometrics
and healthcare to video surveillance and security. In
psychological terms, face identification is a process through which humans
locate and attend to faces in a visual scene.
One can
consider face detection as a specific case of object class detection. A
reliable methodology is based on the eigen-face technique and the genetic
algorithm.
Rather
than just simply telling you about the basic techniques, we would like to
introduce some efficient face recognition algorithms (open source) from latest
researches and projects.
- OpenFace
- OpenBR
- SphereFace
- Deep Face Recognition with Caffe Implementation
- Android Face Recognition with Deep Learning
8 from
China, where facial recognition has received the most significant recent
support:
- Dahua: "interest but not adoption"
- Facego: Big parent company, poor marketing
- Hikvision: Downplayed
- Longse: Fac Rec "Just for Show"
- Qualvision: "Frank Comments on NDAA, Face Rec Hype"
- Sunell: Bold Claims
- TVT: "that's gonna piss our customers off"
- ZKTeco: Claims World’s Best Facial Recognition, Calls Hikvision “Cheap Chinese”
Note:
China's most prominent facial recognition providers, SenseTime, Megvii Face++,
and Yitu.
12 outside
of China, mostly US, with one each from Australia, Japan, Russia, South Korea
and Taiwan:
- Avycon: ''It Can Detect A Face"
- Axxonsoft: Frank Comments on 'Accuracy' Ratings
- Ayonix: Emphasis on Speed
- Ever AI: Positions Itself as China/Russia Competitor
- Deepcam: Selling $59 'Facial Recognition' Cams
- Faceron: Obscure Operations
- Geovision: 3D Face Map and Gender Recommendations
- iOmniscient: "20x Cheaper", Touts Chinese Army as Client
- Panasonic: 'We Beat NEC'
- Real Networks / SAFR: US Based Facial Recognition for Schools Solution
- Tough Dog: Tough Time Selling Face Rec Solutions
- Virdi: World's Best / No Evidence
Suprema has a facial recognition
reader called facestation 2 and also had a new face lite which was introduced in
2019. Idemia VisionPass uses visual cameras, IR, and '3D' Time-of-Flight
sensors to establish face 'liveness' and scan faces to verify users. VisionPass
unit supports IR scanning, capacity for more user templates, and is compatible
with OSDP. Idemia pricing is higher, often 2X to 3X higher for VisionPass
compared to facial recognition models emerging from China.
1- There are two main technologies for Face
Recognition:
- Optical solutions (CCTV based): these
are based on algorithm/pixel performance only. It can be used as black listing
(Stadium, Retail, Vandalism) but it is not enough for white listing (i.e.:
access control).
- Infra Red Solutions (Suprema and
others): these are based on Light emission + IR sensors + Algorithm +
processing power. Advantage of IR are: distance 15cm to 1.5m (it filters
background and all related issues), works in any light conditions (unlike CCTV
that can take a face with sun from the side), makeup/painting on face, Face Face/Images
detection (easier than Optical). These ones are safe enough to be used for
white listing (=access control).
2- FaceLite is working same as Suprema
FaceStation2, with Infra-Red templates (it's compatible).
Cool stuff: FaceLite is 43% smaller (size) than
FaceStation2, and the price follow the same 43% off trend. That brings the
Flite IR face recognition reader to the price of Fingerprint reader (= BioLite
Net : BLN2-OAB). But still you have the high performance/reliability/security.
No sacrifice on this!
Limitation: Face template is too big to be encoded
on a card (>8KB) and Suprema Face Models are "evolutive" (maching
learning: each time you check your face on a reader, it is updated). The
related drawback is that Face cannot be stored on RFID cards (EV1/EV2 / Seos).
Instead it is stored in Central Database or in Reader itself (my preference).
The # of face models are limited to 3,000 (1:N, Identification) and to 30,000
(1:1, Verification, that case you need to swipe a card or input an ID before authentication).
Compared to FaceStation2 (FS2), you also lose the second optical camera (that I
like for user interface or Picture logs), you lose the large touch screen, you
lose Android OS, you lose the Video Intercom possibility. But that's in line
with the 43% off in price point!
Privacy: Face templates are stored on central
server (encryption: AES 256) or on readers (AES 128), with possible
"Tamper secure" option => if the reader is removed from wall, it
factory resets and loses all memory (Users, Face Models, Logs, Encryption keys,
..). Face Models are being transported from Central Server <=> Readers
via TCP, using TLS 1.2 encryption/certificate.
Product Name
|
FaceStation 2
|
FaceLite
|
||
Model Name
|
FS2-D
|
FS2-AWB
|
FL-DB
|
|
RFID
|
RF Option
|
125kHz EM & 13.56MHz MIFARE,
MIFARE Plus, DESFire/EV1, FeliCa
|
125kHz EM, HID Prox & 13.56MHz
MIFARE, MIFARE Plus, DESFire/EV1, FeliCa, iCLASS SE/SR/Seos
|
125kHz EM & 13.56MHz MIFARE, MIFARE
Plus, DESFire/EV1, FeliCa
|
Mobile Card
|
NFC
|
NFC, BLE
|
NFC, BLE
|
|
Protection
|
Ingress Protection
|
Not supported
|
Not supported
|
Not supported
|
Vandal Proof
|
Not supported
|
Not supported
|
Not supported
|
|
Face
|
Template
|
SUPREMA
|
SUPREMA
|
SUPREMA
|
Extractor / Matcher
|
SUPREMA
|
SUPREMA
|
SUPREMA
|
|
Live Face Detection
|
Supported
|
Supported
|
Supported
|
|
Capacity
|
Users (1:1)
* Based on one face enrollment per user |
30,000
|
30,000
|
30,000
|
Users (1:N)
* Based on one face enrollment per user |
3,000
|
3,000
|
3,000
|
|
Max. Face Enrollment per User
|
5
|
5
|
5
|
|
Text Log
|
50,00,000
|
50,00,000
|
50,00,000
|
|
Image Log
|
50,000
|
50,000
|
Not supported
|
|
HW
|
CPU
|
1.4 GHz Quad Core
|
1.4 GHz Quad Core
|
1.2GHz Quad Core
|
Memory
|
8GB Flash + 1GB RAM
|
8GB Flash + 1GB RAM
|
8GB Flash + 1GB RAM
|
|
LCD Type
|
4” color TFT LCD
|
4” color TFT LCD
|
2” color TFT LCD
|
|
LCD Resolution
|
800 x 480 pixels
|
800 x 480 pixels
|
320x240 pixels
|
|
Sound
|
24 bit Voice DSP (echo cancellation)
|
24 bit Voice DSP (echo cancellation)
|
24 bit Voice DSP
|
|
Operating Temperature
|
-20°C ~ 50°C
|
-20°C ~ 50°C
|
-20°C ~ 50°C
|
|
Storage Temperature
|
-40°C ~ 70°C
|
-40°C ~ 70°C
|
-40°C ~ 70°C
|
|
Operating Humudity
|
0% ~ 80%,
non-condensing |
0% ~ 80%,
non-condensing |
0% ~ 80%,
non-condensing |
|
Storage Humidity
|
0% ~ 90%,
non-condensing |
0% ~ 90%,
non-condensing |
0% ~ 90%,
non-condensing |
|
Weight
|
Device: 548g
Bracket: 74g (Including washer and bolt) |
Device: 548g
Bracket: 74g (Including washer and bolt) |
Device: 296 g
Bracket: 41 g (Including washer and bolt) |
|
Dimension (WxHxD, mm)
|
141 x 164 x 125
|
141 x 164 x 125
|
80 x 160.3 x 71.8
|
|
Tamper
|
Supported
|
Supported
|
Supported
|
|
Interface
|
Wi-Fi
|
Not supported
|
Built-in, IEEE 802.11 b/g
|
Not supported
|
Ethernet
|
10/100/1000 Mbps, auto MDI/MDI-X
|
10/100/1000 Mbps, auto MDI/MDI-X
|
10/100 Mbps, auto MDI/MDI-X
|
|
RS-485
|
1ch Host or Slave (Selectable)
|
1ch Host or Slave (Selectable)
|
1ch Host or Slave (Selectable)
|
|
Wiegand
|
1ch Input, 1ch Output
|
1ch Input, 1ch Output
|
1ch Input or Output (Selectable)
|
|
TTL Input
|
2ch Inputs
|
2ch Inputs
|
2ch Inputs
|
|
Relay
|
1 Relay
|
1 Relay
|
1 Relay
|
|
USB
|
USB 2.0 (Host)
|
USB 2.0 (Host)
|
USB 2.0 (Host)
|
|
SD Card
|
Not supported
|
Not supported
|
Not supported
|
|
PoE
|
Not supported
|
Not supported
|
Not supported
|
|
Intercom
|
Supported
|
Supported
|
Not supported
|
|
Electrical
|
Power
|
Voltage: DC 24 V
Current: Max. 2.5 A |
Voltage: DC 24 V
Current: Max. 2.5 A |
Voltage: DC 24 V
Current: Max. 2.5 A |
Platform
|
BioStar 2
|
Supported
|
Supported
|
Supported
|