Face in Aadhaar EcoSystem - Supported Documents and Integration Files for Android and Windows OS

472 views
Skip to first unread message

Suraj Nair

unread,
Jun 26, 2018, 2:13:09 AM6/26/18
to Aadhaar Registered Devices Discussion Group
Dear All;

Please find below link for downloading the documents and integration files for Android and Windows OS, provided as a reference to device provider for Integrating Face as a modality in their respective RD service.

For Android 
-----------
1. Introduction_Face-Detection-In-RD-Service.pdf
2. UIDAI_FACE_LIB-RELEASE.aar
3. UIDAI-FACE-LIB-USER-MANUAL.pdf
4. Link for downloading is mentioned below :-
      


For Windows 
-----------
1. Introduction_To_Face-Detection_Flow_Using_OPENCV.pdf
2. OPENCV-FACE_DETECTION_FUNCTON.pdf
3. OPENCV_Details<Folder>
   a.Cascade Classifier XML 
          - eyes.xml
  - haarcascade_frontface_alt.xml
4. TEMPLATES <Folder>
    - Close_eye_template.jpg
    - Open_eye_template.jpg
5. Link for downloading is mentioned below :-
      
   

DISCLAIMER
------------------- 
The sample code or utility or tool described herein is provided on an "as is" basis, without warranty of any kind. UIDAI does not warrant, guarantee or make any representations of any kind with respect to the sample code and does not make any representations or warranties regarding the use, results of use, accuracy, timeliness or completeness of any data or information relating to the sample code. UIDAI disclaims all warranties, express or implied, and in particular, disclaims all warranties of merchantability, fitness for a particular purpose, and warranties related to the code, or any service or software related thereto. 
     
UIDAI is not responsible for and shall not be liable directly or indirectly for any direct, indirect damages or costs of any type arising out of use or any action taken by you or others related to the sample code. This is released as a reference implementation and in case of using our reference implementation, license conditions are to be adhered to and UIDAI will not be responsible for any violation. 

                                         THIS IS NOT A SUPPORTED SOFTWARE.

Mahesh Patel

unread,
Jun 26, 2018, 2:35:26 AM6/26/18
to aadha...@googlegroups.com

Dear surajji,

 

1. Windows: I have referred and implemented your algorithm xmls for eye, but I have doubt that, in your xml, you are detecting both eyes. But what happened if person having only one eye?

This is general question, because yet I have not tested it with actual case where only one eye is present.

 

2. Android: In android, you have suggested “Google Vision API”. This is possible with Standard RD service but what about non-standard Rd service where device may not having google play service?

 

 

Regards,
Mahesh Patel

Mantra Softech India Pvt Ltd
ISO 9001:2008 Certified Company

Network: AHMEDABAD–BANGALORE–HYDERABAD–CHENNAI-DELHI–PUNE–MUMBAI–JAIPUR

--
You received this message because you are subscribed to the Google Groups "Aadhaar Registered Devices Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aadhaar_rd+...@googlegroups.com.
To post to this group, send email to aadha...@googlegroups.com.
Visit this group at https://groups.google.com/group/aadhaar_rd.
To view this discussion on the web visit https://groups.google.com/d/msgid/aadhaar_rd/5bc4ca57-a17e-4308-8d6e-831f85dee22c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Pavel Cuchriajev

unread,
Jun 26, 2018, 2:59:40 AM6/26/18
to aadha...@googlegroups.com

Hi,

 

I have a quick question:

 

  1. Is face capture part of the trusted environment?
  2. Is there an official change of the spec (again referring to the first question)? there should be formal modification of Registered Devices spec to solve all of the arising questions.

 

BR,

Pavel

Netaji Rao

unread,
Jun 27, 2018, 3:31:51 AM6/27/18
to aadha...@googlegroups.com
Hi Team,

We would need updated RD specification doc with face auth requirements and changes.

Thanks,
Netaji Rao D

Ketan Upadhyay

unread,
Jun 27, 2018, 5:28:58 AM6/27/18
to aadha...@googlegroups.com

Dear Netaji

For L0 registered device , the specs change may not be required. As the document already mentions about face and working with biometrics. Face support is to be provided by RD of finger or Iris and hence calling intent and method will remain same. The face capture should be in Native code (for windows and linux, in line with biometrics) and for android using in-built camera

 

Regards

Ketan Upadhyay

+91 9327208489

EMAIL SIGN

 

From: aadha...@googlegroups.com <aadha...@googlegroups.com> On Behalf Of Netaji Rao
Sent: 27 June 2018 13:02
To: aadha...@googlegroups.com
Subject: Re: [aadhaar_rd] Face in Aadhaar EcoSystem - Supported Documents and Integration Files for Android and Windows OS

 

Hi Team,

 

We would need updated RD specification doc with face auth requirements and changes.

Thanks,
Netaji Rao D

 

On Tue 26 Jun, 2018, 12:29 PM Pavel Cuchriajev, <pavel.cu...@nextbiometrics.com> wrote:

Hi,

 

I have a quick question:

 

1.       Is face capture part of the trusted environment?

2.       Is there an official change of the spec (again referring to the first question)? there should be formal modification of Registered Devices spec to solve all of the arising questions.

Netaji Rao

unread,
Jun 27, 2018, 5:35:21 AM6/27/18
to aadha...@googlegroups.com
Dear Ketan,

Does it mean, face capture is mandatory with each fingerprint/iris auth? I guess, Application is supposed to request for face (optional) through PidOptions.

Also, who is supposed to sign (bs) face biometrics? I think, Fingerprint or Iris vendor cannot guarentee/sign an image captured outside their device/SDK.

Thanks,
Netaji Rao D

Ketan Upadhyay

unread,
Jun 27, 2018, 5:43:13 AM6/27/18
to aadha...@googlegroups.com

Dear Netaji

 

The points you had mentioned are discussed in last meeting. The BS is to be signed by Finger or Iris device provider key. The Face support is mandatory to be added to all currently certified RD. The face image is not captured outside RD, but by RD.

 

“Also, who is supposed to sign (bs) face biometrics? I think, Fingerprint or Iris vendor cannot guarentee/sign an image captured outside their device/SDK.” :- UIDAI is not allowing only face authentication and supported by other biometrics.

Pavel Cuchriajev

unread,
Jun 27, 2018, 5:44:19 AM6/27/18
to aadha...@googlegroups.com

Hi,

 

Exactly, especially in case of Windows, using filters and pipeline can be changed dynamically. Also current spec does not mention NOTHING about face capture and where it should occur. What Ketan wrote is just assumption, but unless it is in paper – I don’t think it is wise to ask all vendors to comply with it.

 

BR,

Pavel

Pavel Cuchriajev

unread,
Jun 27, 2018, 5:45:10 AM6/27/18
to aadha...@googlegroups.com

Capture by RD and capture in trusted environment differs. Especially in case of future L1 😊

 

BR,

Pavel

Netaji Rao

unread,
Jun 27, 2018, 5:50:18 AM6/27/18
to aadha...@googlegroups.com
Face support will be added by RD, but does every auth require face also in the PID block? Does it supposed to be opted by Application through PidOptions?

Thanks,
Netaji Rao D

Ketan Upadhyay

unread,
Jun 27, 2018, 5:50:39 AM6/27/18
to aadha...@googlegroups.com

Yes, L1 is differs and it should be diff hardware. For L0 the doc mentions about biometrics without any diff between Finger/Iris/face. And option does mention about pcount (face)

 

This is my opinion and anything in particular needs to be amended is last specs, we should suggest to UIDAI team with exact details.

 

Regards

Ketan Upadhyay

+91 9327208489

EMAIL SIGN

 

sachin

unread,
Jun 27, 2018, 5:50:48 AM6/27/18
to Aadhaar Registered Devices Discussion Group
Hi, 
Can you please suggest the Open CV version you are using  for integration of the above SDK in Windows platform
Any appropriate links for samples would be very helpful for us in implementing this faster.

Regards,
Sachin N

Netaji Rao

unread,
Jun 27, 2018, 5:58:03 AM6/27/18
to aadha...@googlegroups.com
Ok, if pcount is 1, then capture face and add it to pidblock, otherwise no-face capture.

Any guidelines on face image size/compression ratio?

Thanks,
Netaji Rao D

Ketan Upadhyay

unread,
Jun 27, 2018, 6:03:12 AM6/27/18
to aadha...@googlegroups.com

“Face support will be added by RD, but does every auth require face also in the PID block? Does it supposed to be opted by Application through PidOptions?” yes, it is to be decided by application.

Face image size and compression, the UIDAI team has mentioned about max 30% quality compression with color photo. For B&W photo, they will revert. Face should be front facing, the size I do not remember, I was asked in meeting, but I missed.

Ketan Upadhyay

unread,
Jun 27, 2018, 6:07:00 AM6/27/18
to aadha...@googlegroups.com

Hi

Suraj from UIDAI team has shared these details.

 

Ketan Upadhyay

+91 9327208489

EMAIL SIGN

 


Sent: 27 June 2018 15:21
To: Aadhaar Registered Devices Discussion Group <aadha...@googlegroups.com>

--

You received this message because you are subscribed to the Google Groups "Aadhaar Registered Devices Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aadhaar_rd+...@googlegroups.com.
To post to this group, send email to aadha...@googlegroups.com.
Visit this group at https://groups.google.com/group/aadhaar_rd.

Netaji Rao

unread,
Jun 27, 2018, 6:07:37 AM6/27/18
to aadha...@googlegroups.com
That's true Pavel, Android integrated camera is trustee to some extent as safetynet etc checks are in place, but a USB camera on Windows?

Thanks,
Netaji Rao D

Manish Nandaniya

unread,
Jul 2, 2018, 6:37:49 AM7/2/18
to Aadhaar Registered Devices Discussion Group
Hi Mr. Suraj Nair,

I am authenticating with face support integrated to RDService (BioEnable Technologies Pvt. Ltd.) using latest "Aadhaar Test App" (PoC) and getting below responses:

1. PidOptions: fcount=1,ftype=0,pcount=1,ptype=0, remaining standard., with Aadhaar: own staging aadhaar, Finger: own, Face: own, Auth Response: success.

2. PidOptions: fcount=1,ftype=0,pcount=1,ptype=0, remaining standard., with Aadhaar: own staging aadhaar, Finger: own, Face: other's, Auth Response: success.

3. PidOptions: fcount=1,ftype=0,pcount=1,ptype=0, remaining standard., with Aadhaar: own staging aadhaar, Finger: other's, Face: own, Auth Response: 300, biometric did not match.

4. PidOptions: fcount=1,ftype=0,pcount=1,ptype=0, remaining standard., with Aadhaar: own staging aadhaar, Finger: other's, Face: other's, Auth Response: 300, biometric did not match.

5. PidOptions: fcount=0,ftype=0,pcount=1,ptype=0, remaining standard., with Aadhaar: own staging aadhaar, Finger: NA, Face: own, Auth Response: 931, UIDAI internal error.

(Note: Used hardcoded PidOptions for authentication via Aadhaar Test PoC Test App)

With above test cases I am concluding, adding face bio tag <Bio type="FID" posh="FACE" bs="...bs...">...encoded biometric...</Bio>  in Pid block is resulting in no effect, when testing using "Aadhaar Test App" (PoC). (Please correct for any invalid test case.)

Also, as I am testing using "Aadhaar Test App" (PoC), is there any dedicated or updated PoC test app for face authentication that can be used to ensure whether face support has added in RDService correctly?

Kindly suggest for the problem.

Regards,
Manish Nandaniya
Android Developer
BioEnable Technologies Pvt. Ltd. Pune

aggarwa...@gmail.com

unread,
Jul 2, 2018, 7:33:16 AM7/2/18
to Aadhaar Registered Devices Discussion Group
Hi

Is there any specific version of opencv which we have to use in RD Services for windows ?, 

In Android 

if (Facedata != null) {
String[] dataToSend = new String[2];
dataToSend[0] =Facedata[0];
dataToSend[1] = Facedata[1];
Message msg = new Message();
msg.what = 2;
msg.obj = dataToSend;
*/Your Service/*.mMyServiceHandler.sendMessage(msg);


Is the  Facedata[1] face Image in base64 format ?
or Template of face image  in base64 format ?



Regards
Vishnu Aggarwal



On Tuesday, June 26, 2018 at 11:43:09 AM UTC+5:30, Suraj Nair wrote:

ashish.s...@iclimbsystems.com

unread,
Jul 3, 2018, 3:15:49 AM7/3/18
to Aadhaar Registered Devices Discussion Group
It will be helpful if we can get sample XML of success case request / response which includes fingerprint as well as face ID. Should resolve many of the queries.
Thanks !

Regards,
Ashish

Sridhar(Eydentify)

unread,
Jul 4, 2018, 7:15:23 AM7/4/18
to aadha...@googlegroups.com

Dear All,

 

As mentioned in the document for Windows Face Authentication using OPENCV, we are able do the liveness check and capture the image.

In android we are getting FID template from the Library, there is no API in OPEN CV to get the FID Template in windows.

Kindly suggest how to get the Face Template in Windows.

 

Regards,

 

Sridhar Pashya

 

 

From: aadha...@googlegroups.com <aadha...@googlegroups.com> On Behalf Of Suraj Nair
Sent: 26 June 2018 11:43
To: Aadhaar Registered Devices Discussion Group <aadha...@googlegroups.com>
Subject: [aadhaar_rd] Face in Aadhaar EcoSystem - Supported Documents and Integration Files for Android and Windows OS

 

Dear All;

--

aggarwa...@gmail.com

unread,
Jul 4, 2018, 7:18:33 AM7/4/18
to Aadhaar Registered Devices Discussion Group
Hi Sridhar

Please tell that which version of OpenCV is used  ?

Regards

Sridhar Pashya

unread,
Jul 4, 2018, 7:22:00 AM7/4/18
to Aadhaar Registered Devices Discussion Group
Hi Vishnu ,

OPENCV version 3.4.1

Regards, 

Sridhar Pashya 
Sent: Wednesday 4 July 2018 4:48 PM
To: Aadhaar Registered Devices Discussion Group
Subject: Re: [aadhaar_rd] Face in Aadhaar EcoSystem - Supported Documents and Integration Files for Android and Windows OS

aggarwa...@gmail.com

unread,
Jul 4, 2018, 7:23:55 AM7/4/18
to Aadhaar Registered Devices Discussion Group
Hi

I had also treid but this version not have 32 bit library

Regards
To post to this group, send email to aadh...@googlegroups.com.

prasadk

unread,
Jul 11, 2018, 6:09:15 AM7/11/18
to Aadhaar Registered Devices Discussion Group
We are also facing issue while extracting FID template from face image ..
we are using opencv ver 3.4.1,
able to use haarcascade_eye.xml to detect Eyes & haarcascade_frontalface_default.xml to detect Face,
but not found any OPENCV API to extract FID Template from face image.

pls suggest.


On Tuesday, June 26, 2018 at 11:43:09 AM UTC+5:30, Suraj Nair wrote:
Reply all
Reply to author
Forward
0 new messages