The Iris Recognition SDK

Swallow Incubate Co., Ltd. utilizes Panasonic patent technology and
we are the only company in Japan to research and develop「distinctive Iris Recognition engine」.

Structure of The Iris Recognition system

Currently, most of the commercialized 「Iris Recognition system」in the world use 「Daugman algorithm」suggested by professor Daugman at US/Cambridge University in 1992.
In brief description of the mechanism, it generates personal 「iris code」using photographed 「iris image」, matches it to the databased 「iris code」, and determine it as the same person or not.
We aim to materialize 「the improvement of recognition accuracy」of our 「Iris Recognition system」by deriving 「our distinctive algorithm」from the Daugman algorithm.

Flow of our Iris Recognition system

The processing flow of our developing 「Iris Recognition system」has below 8 steps.

  1. Detecting eyes.It detects personal eyes out of the photographing objects by utilizing the OpenCV library.
  2. Focusing.Correctly focus (focusing degree) the camera to take picture of iris pattern with resolution that can fulfill the recognition.
  3. Photographing the eye images using an infrared camera.It is difficult to take pictures of the iris patterns of eyes because of specular reflection on the cornea, so use an infrared camera and infrared light.
  4. Extracting the iris image from eye image.Detect the pupil and outer edge of the eye image, detect the iris outside of the pupil, the scleral boundary of outside, and then extract the iris image.
  5. Converting the iris image to the iris code.Binarize in fixed length the iris image extracted , and generate the iris code.
  6. Saving the iris code into the database.Save the iris code into the database. It is also possible to save the personal multiple iris codes as a link.
  7. Deletion of the iris images.There is no leak risk because the iris images are deleted right after photographing them.
  8. Executing Iris Recognition.At the recognition time, the iris image is converted to iris code and search the iris code with high similarity within the threshold value out of the database, and then the recognition result is returned.

The Iris Recognition accuracy

The accuracy of 「the Iris Recognition」enables to balance 「FAR」and「FRR」, and adjust the threshold value of the variance rates (hamming distance).

・About FAR

「FAR(False Acceptance Rate)」showing the level of security is the probability to wrongly accept an unregistered person's iris as a registered person's one.
In the authentication industry, the lower 「FAR」 is better.

・About FRR

「False Rejection Rate (FRR)」is the probability to wrongly reject a registered person as an unregistered one.
In the authentication industry, the lower 「FRR」is also better as same as 「FAR」.

In the authentication process, it calculates 「the variance rates(hamming distance)」when 「both iris codes」 are compared.
It's mechanism determines 「accept or deny」by setting up a certain 「threshold value」 on its 「variance rates(hamming distance)」.
In 「The Iris Recognition system」, 「The accuracy of recognition」can be adjusted by where to set up the variance rate(hamming distance). To set up 「FAR」to more sever level means the threshold value of 「variance rates(hamming distance)」is set up lower level but 「FRR」gets higher level, and the reverse situation as well.

Relations between FAR and FRR

◎ certify

FAR Down
FRR Up

× denial

FAR Up
FRR Down

  • difference
    degree 0%
  • difference degree20%
  • difference degree100%

When to install 「The Iris Recognition system」, the important thing is to tune it to 「appropriate recognition accuracy」to fit to the usage scene.

Problems of the Iris Recognition

There are 3 below major problems to overcome to make 「The Iris Recognition」practical use.

① 「loss of iris pattern」by the external light glare and the reflection of the infrared light.

For the Iris Recognition influenced by the lights from outside like using the Iris Recognition at outdoor and inside with large windows, loss of「iris pattern」is seen by the glare of the lights from outside.
In case of using glasses and sunglasses, also 「the iris pattern」gets loss because of the reflection of 「the infrared light」.

External light glare

External light glare

Infrared light reflection

Infrared light reflection

Glare to the iris area

Glare to the iris area

Loss of iris pattern

② 「The pupil diameter change」by the「brightness condition」change

Normally we are not aware, but our pupil gets smaller under bright environment and our iris get bigger. Under dark environment the pupil gets larger and the iris gets smaller.
Due to the「iris area」change by the「brightness condition」change , the problem of rejection of the person in question occurs.

Under bright environment pupil small/iris large

Under bright environment
pupil small/iris large

Under bright environment pupil small/iris large
Under normal environment pupil middle/iris middle

Under normal environment
pupil middle/iris middle

Under normal environment pupil middle/iris middle
Under dark environment pupil large/iris small

Under dark environment
pupil large/iris small

Under dark environment pupil large/iris small

Change of the iris area

③ Expensive Iris Recognition module

Current commercialized 「Iris Recognition modules」are expensive, and the retail price will get high through using those expensive modules for production and sale.
This is one of the reasons of 「the Iris Recognition is expensive(price/cost)」.

GEMINI-CORE

SDK
camera module
About $400〜$600

Both eyes iris photography camera module

Both eyes iris
photography camera module
About $480

One eye iris photography camera module

One eye iris
photography camera module
About $190

Overcome of the Iris Recognition problems and the strengths of our Iris Recognition technology

To solve above problems of ① 「loss of iris pattern」by the external light glare and the reflection of the infrared light and ②「The pupil diameter change」by the「brightness condition」change, we are developing with utilizing 「Panasonic patent technology」.

① 「loss of iris pattern」→ 「iris composition technology」

This technology supplements by composing lossed iris pattern by the external light glare and the reflection of lens.

iris
iris
iris

Compose the lossed iris by the glare and supplement

② 「The pupil diameter change」→ 「Iris multi-template method」

To increase the recognition accuracy in outside and inside where the brightness conditions are different, save multiple iris codes which have different pupil diameters.

iris
iris
iris

Increase the recognition accuracy by saving multiple iris codes which were photographed at outside and dark place.

Utilizing patent technologies

  • Patent number JP3586456: Personal identification method and Personal identification device
  • Patent number JP3639291: Personal identification method, Iris registration device, Iris Recognition device, and Personal identification program

Also, as the solution of above problem 「③ Expensive Iris Recognition module」, we recommend 「SDK license」of 「the Iris Recognition engine」which we are developing.

③ 「Expensive Iris Recognition module」→ 「SDK license of the Iris Recognition engine」

The license of SDK itself enables to develop the Iris Recognition module, flexible integration・development of various products, and it's available to adopt flexible license system.

Our company

Our company

Development/Sales company

Development/
Sales company

PC_tablet_smartphone VR ATM

About the Iris Recognition SDK

「SDK」stands for 「Software Development Kit」, and it's the package for development.
It is possible to integrate our 「Iris Recognition SDK」 into various applications and engines.
The major merits of using our 「Iris Recognition SDK」are below.

① Possible to combine with other engines, AI, and IoT services

The merit of 「SDK」is flexibility to combine with other functions. It is also possible to install it into the existing products, and the products / services under development.

Installation of the Iris Recognition / examples of utilization

  • Entrance and exit / Admission controlEntrance and exit / Admission control
  • Smartphone and PCSmartphone and PC
  • Electronic payment / CryptocurrencyElectronic payment / Cryptocurrency
  • Automobile / Connected carAutomobile / Connected car
  • AI / RobotAI / Robot
  • VR productsVR products
  • Attendance management systemAttendance management system
  • Bank / ATMBank / ATM
  • SafeSafe
  • ApartmentApartment

② Available to develop services with own brand

Our 「Iris Recognition SDK」is one of factors in your developing services, so it is not required to display our company name.

③ Accelerate your businesses by utilizing patent technology

We hope 「Panasonic patent technology」researched and developed which spent a large capital will be used effectively by many customers. By using the advanced technology 「Iris Recognition SDK」, we support to increase your brand value, product value, and good start of your new business development.

Comparison with other 「Iris Recognitions」

We are currently the only company to research and develop 「the Iris Recognition technology」in Japan. 「The Iris Recognition mounted products」used in Japan have installed 「The Iris Recognition engine」under overseas technology.
We research and develop the「Iris Recognition engine」in Japan for practical use to be able to license provide the 「SDK itself」which can easily be used for product/service development.

The Iris Recognition system/Product comparison

IrisAccess series iCAM series

IrisAccess series
iCAM series

Company nameIRIS ID, Inc.

Using algorithmDaugman algorithm

Features

  • The Iris Recognition research and development division of LG Electronics Incorporated (Korea) was spinout.
  • Price approx. $2500/unit
Active IRIS

FPC Active IRIS

Using algorithmDaugman algorithm

Features

  • World-first Iris Recognition smartphone in Japan (Fujitsu Arrows)
  • FingerPrint Cards merged DeltaID in 2017.
EF-45

EF-45

Company nameCMITech Company Ltd.

Using algorithmDaugman algorithm

Features

  • Hybrid device of face recognition x Iris Recognition (face recognition is option)
Gemini Camera

Gemini Camera

Company nameIritech, Inc.

Using algorithmDistinctive algorithm

Features

  • Provide the iris photographing module
  • Module $400〜600/each
  • Difficult to provide the recognition SDK itself
BM-ET200

BM-ET200

Company namePanasonic, Inc.

Using algorithmDaugman algorithm

Features

  • Distinctive Iris Recognition device
  • Currently suspension of the sale
The Iris Recognition technology SDK

The Iris Recognition technology SDK

Company nameSwallow Incubate Co., Ltd.(our company)

Using algorithmDistinctive algorithm

Features

  • Utilize Panasonic patent related to the iris authentication
  • Available to provide the recognition SDK itself
  • Flexible license fee

Introduction of our 「Iris Recognition engine」

Eyepiece type Iris Recognition demo

Eyepiece type Iris Recognition demo

Eyepiece type Iris Recognition demo device

Eyepiece type Iris Recognition demo device

Short distance type (~ 30cm) Iris Recognition demo device

Short distance type (~ 30cm) Iris Recognition demo device

Prototype module

Prototype module

Swallow Incubate is the only company in Japan who is working on the practical use development of 「The Iris Recognition engine」. If you are interested in, please contact us.

Top