Introducing Face++ Cognitive Services: Another Step Towards Visual Intelligence

2017-01-13

Dear Face++ users,

We are happy to introduce to you a new product - Face++ Cognitive Services, a platform offering leading, deep learning-based image recognition APIs & SDKs with free and premium options on your demand. Follow the link below to 

Background

We established our vision as Power Human with AI ever since the first version of Face++ launched nearly five years ago. We have been dedicated to deliver leading, reliable service to every developer and customer. Fortunately, our effort has paid off. Since Face++’s debut, thousands of users have employed Face++ facial recognition APIs and SDKs, making Face++ the most popular platform in the market. We also made great achievements in facial recognition research field, winning several world-class competitions including FDDB, 300-W and LFW.

As we continue to expand our capability set of computer vision available to Face++ users, we challenged ourselves to find a polished way to deliver service for developers and businesses of all sizes, and now we’re ready to unveil what we’ve been working on.

What’s New?

Face++ Cognitive Services is an all new platform. To support its features, we ultimately rebuilt the platform from the ground up, with our prestigious computer vision algorithms at the core, even faster, simpler, and better.

Most Up-to-date Algorithms

All the APIs and SDKs offered by Face++ Cognitive Services are powered by up-to-date algorithms, which are much more efficient, accurate and robust. We goes far beyond the old version of Face++, and once again lead the market of facial recognition technologies. For example, based on our internal test result, the response time of new Face Detect API is 65% shorter than the old version, and the response time of Analyze API is 69% shorter. You can test the performance of our service on your own next week. — It’s completely free to start.

Better Product, Better Service

We’ve made great effort to develop a better product with simplicity and reliability. Besides the re-designed interface and profound demonstration, Face++ Cognitive Services platform provides a complete console dashboard and documentation center, which enable you to explore every detail of our products, to scale and view your usage and billing. We also offer SLA and support plans for premium options.

More than Facial Recognition

The name Face++ Cognitive Service echoes our initial dream of Visual Intelligence. While dedicated to offer the best facial recognition technologies, we exploit our advantage of artificial intelligence and expand the capability set to include image recognition, OCR, object recognition, etc. These capabilities are built on the same deep learning technology, with leading accuracy and performance. We shall officially launch new capabilities next month.

Free & Premium Options

We’re committed to meet the requirements of every customer. Face++ Cognitive Services offers both free and premium options for all users. You can get free access to all Face++ APIs, with no usage limits, no upfront fees, no commitments. And we offer premium options including Pay-As-You-Go and Monthly Plan, for businesses of all sizes, meeting the demand of higher performance and technical support. There are two featured licensing plans for SDKs as well. The premium service shall be launched next month.

About Old Face++

Can I Continue to Use the Old Version?

The old Face++ website has been switched to old.faceplusplus.com after the launch of Face++ Cognitive Services. The old version of API service will remain running for a certain month. However, every user is highly recommended to use the all new Face++ Cognitive Services, since we will gradually shut down the old service in the following months.

Can I Use Same Account and API Credentials of Old Face++ on New Version?

Unfortunately you cannot use the same account and API credential of old version in Face++ Cognitive Services. Face++ Cognitive Services is an all new platform, with rebuilt website and APIs. You need to first register on our website, get your API Key and check the API Reference. We’ve made the starting process simple and easy. You can start building in a few clicks.

We’ve been greatly motivated by positive feedbacks and encouragements from our customers, and we look forward to continuing to serve you. If any problem, please do not hesitate to contact us at support@faceplusplus.com.

Thank you,

Face++ Team

Important! APP release upgrade guide

2014-04-03

We are excited to know that you’re integrating Face++ into your app! In order to maintain a smooth and stable service, we are now set up some rules regarding the proper API usage.

  1. In development status, your app may contain at most 5 Facesets/Groups and 100 Persons. If you’d like to lift the restrictions, you’ll need to upgrade your app to release status.
  2. When applying for release status upgrade, you need to provide valid app information and app link. Your request will be handled in two work-days.
  3. Release status app created before Nov. 2013 without detailed app information should provide them as soon as possible. Please go to Developer/My App/Manage Now/Edit Profile, edit it and then submit those info. Please NOTE that if you fail to complete these steps before July 6, your app will fall back to development status. 

Thanks for your long-time supporting.

Face++ achieved three No.1 performance on world-class tech evaluations!

2014-03-12

Face++ is on his road being the world's No.1  facial technology company. Nowadays, some reports have been released by relevant institutions. Face++ has won the first place in face detection, landmark detection and face recognition competitions.

  1. Face++ delivers state-of-art face detection performance in FDDB. See Face++ detect performance report.
  2. Face++ team won the “300-Face in the wild” challenge during the workshop of ICCV2013. See Face++ in ICCV2013.
  3. Face++ team got the champion of LFW with accuracy of 97.27% , beat Facebook team (Former Face.com team) of 97.25% . Click here to download the paper.

All the new techs will be coming soon as API service!

Want to get more Face++ news? Follow us on Facebook & twitter!

Face++ detect performance report

2014-02-24

According to the FDDB (Face Detection Data Set and Benchmark) report, Face++ delivers state-of-the-art face detection performance among its commercial competitors.

FDDB is a data set of face regions designed for studying the problem of unconstrained face detection. This data set contains the annotations for 5171 faces in a set of 2845 images taken from the Faces in the Wild data set.

In the below diagrams, the horizontal axis represents the number of false positive faces reported by the detector, while the vertical one shows the true positive rate. The closer the curve approaches to the upper-left corner, the better the corresponding detector does on this benchmark.

dicRoc
contRoc

It can be seen from the discrete score diagram, Face++ holds the best performance 0.8 TPR (True positive rate) with nearly 30 false positives. Notice that, in continuous score situation, Face++ detector enjoys a great performance margin over other detectors.Click here to get more.

Face++ delivers the leading face techs to global developers. It’s our pleasure to have your witness of improvements of Face++. Stay focus!

Face++ in ICCV2013 (Ⅱ)

2013-12-12

Face++ researchers have just returned from ICCV 2013, Sydney. It was an exciting adventure for the team!

During the conference, demos exhibited by Face++ received tremendous attention from both the academic and industrial community. Face++ team had a lot of interesting discussions with the researchers on how to help advance face technology from industry side. Moreover, many professionals with computer vision background showed great interest in applying Face++ service into their own products and applications.

Another piece of good news is that Face++ team won the "300-Face in the Wild" challenge during the workshop on Dec 7. This is the first facial landmark detection challenge in the "wild" scenario, targeting to motivate the progress in this direction. By leveraging the recently developed "deep learning" technology, Face++ team successfully proposed a coarse-to-fine framework for the problem, which has a great advantage over traditional shape-model based methods. This technology is expected to be released on the cloud platform soon! 

"New logo, new service". Face++ team continues to provide leading face technology service to the community, and works closely with developers to apply those technologies into new problems and areas. Keep in touch with us and glad to hear from you!

dsc_3961

 

Face++ in ICCV2013 (Ⅰ)

2013-12-04

Face++ team starts a new journey of challenges!

ICCV is the premier international computer vision event. With its high quality, it provides an exceptional value for students, academics and industry researchers. This year, Face++ team joins the big party as one of the main sponsors and exhibitors.

In the conference, many prestigious professors and promising young scholars come to visit Face++. They put forward updated questions and constructive suggestions,  show enormous curiosity and great expectations.

Face++ shows four products during the conference, including "Ad Statistics", "Find Your Girl", "3D-Facial Expression Controller", and Android Face ScreenLocker. Meanwhile, Face++ Research Toolkit (Matlab SDK) is released. Face++ hopes to make a contribution to computer vision community and play an important role to advance face recognition technology.

Stay focus on Face++!

Update Released: 83 key points for one face!

2013-11-25

Hello, Face++ developers. We are pleased to announce that some new features have been released in the latest version API. The Dev Center is also improved in some aspects. New features provided by latest API

  1. A new local face detector has been released (iOS & Android). Developers can download it on the “My App” page. Please refer to Application Management System & Command Line Tool for detailed instructions.
  2.  A new /detection/landmark API is provided. It is able to detect 83 key facial points, including eyebrows, eyes, nose, mouth, and face contour. We hope that it will be a powerful tool for developers to build up interesting interactive applications. Developers can find more details in /detection/landmark.
  3.  A newer version of race analysis algorithm is released. This upgrade reduces the error rate by 50% in our benchmark testing. We will continue to upgrade the back-end API to provide more accurate results for all the face attributes.
Improvements on the website
  1. An overview page of Face++ API is added. It collects all the API interface documents so that developers can quickly locate the information they need.
  2.  “Quick Start” page is reorganized; with step-by-step instructs elaborated to help the fresh users to utilize the advanced face recognition technologies provided by Face++.
  3.  A roadmap for creating new Face++ Apps is added in the Dev Center. Developers can just follow the steps to create their first Face++ App.

Important Update Released: No more fee! More functions!

2013-10-16

Face++ announces that all online API are FREE from now! Developers can call the API as many time as you want! Detailed new policies are,

  1. No fee for online API.
  2. No limit on times of API request.
  3. To prevent evil attack, the number of synchronized API requests will be limited. That is, suppose the number of synchronized API requests is k, then at any time, Face++ server will only process no more than k request from one application. For trial version, k=1. For development version, k=3. For online version, there is no limit on k. Applications created before Oct 16, 2013 are all online version by default (but evil usage will be degraded).
In the latest released version, more functions are added. They are,
  1. New attribute in /detection/detect API:
    • landmark (25 key points for one face)
    • glass (whether the person wear a pair of glass or not)
    • smile (whether the person is smiling or not)
    • pose (the 3D orientation of a face, including Pitch/Roll/Yaw)
  2. Update algorithm for /recognition/search and /train/search: We implemented a new algorithm with better performance for face search. You can call /train/search first and then call /recognition/search to have a try with the new algorithm.
 

Update:Grouping API and Premium Service is launched, and more!

2013-04-05

Grouping of photos by similarity of faces is one of the key features of many photo/album related applications. Based on feedback from many of you, we are pleased to announce Grouping API support. Uses and comments are welcomed! For more information on Grouping API, please visit here.

As more and more developersare joining our parade, Face++ API service matures. After careful review and revision of service policy, all of the current service will continue under a new umbrella called “Face++ Community API Service”, while a new service oriented towards business use of larger scale is provided as “Face++ Premium API Service”. For more information, please visit here.

Apart from above, following release of python, Java, iOS SDKs, Ruby SDK comes up, plus a developer-contributed C# SDK. Buaaparty, thanks!.

Last but not least, we are pleased to announce full HTTPs support added to our API services, while the older way of calling us continue to be supported.

Face++ v1.0 release!

2012-10-25

v1.0-300x117
We are very happy to announce that Face++ v1.0 is released!
We will stick to our promises of API steadiness, meanwhile, expect bugs to fix and performance issues to resolve.
Develop your Apps with Face++ APIs now!
What’s new in v1.0:
1. Finalized API design
2. Added /info APIs
3. Improved performance of Face Verification
4. Improved performance of Face Search
5. Two more attribute in Face Analysis: age & race(asian, black and white)
6. Face++ Command Line Tool release
7. Python SDK released
Any suggestion or questions, please contact us.
attribute_zh-300x159

Google+