Facial Recognition Industry Leaders

Facial Recognition Industry LeadersFacial Recognition Industry Leaders Comparision (and why WE are better)

The artificial intelligence and facial recognition industry is growing rapidly and doesn’t seem to be backing down. As a sector that is reported to be worth anywhere between $7 billion and $15 billion by 2024, it is necessary to understand who the big players are and how they stack up against each other.

We’ve narrowed it down to three big competitors: Kairos, Microsoft Azure Face API and Face++. We have personally tested FaceX’s APIs against those of these 3 giants Facial Recognition Industry Leaders.

But first, let’s introduce our contestants…..

Kairos

Founded in 2012 in Miami, Florida, Kairos is an artificial intelligence startup that focuses on facial recognition and is one of the Facial Recognition Industry Leaders. It has grown quite popular in the industry as one of the best performing facial recognition systems. They offer a range of features such as face detection, face verification, gender identification among others. Although the number of people employed by the company fall below 100, they have been invested in heavily by big players such as American investment bank giants, Morgan Stanley. (Source: Kairos)

Total Funding: $11 million (Source: Crunchbase)

 

Microsoft Azure Face API

Microsoft needs no introduction. Azure, however, is a cloud-based platform that offers a range of service out of which Face API is one. Microsoft Azure is a renowned name among the Facial Recognition Industry Leaders.  They propose many of the standard features along with the addition of emotion detection. Microsoft is a publicly traded company and is why they have no shortage of money coming to propel their research and development.

Face++

Megvii Technology Limited (more popularly known as Face++) is a Chinese artificial intelligence company that is among the largest private firms among Facial Recognition Industry Leaders. Their features presented are similar to that of Microsoft’s Face API. They have some of the most advanced technology in the market considering that they employ just under 5,000 employees and are heavily funded by many of the biggest corporations in China including Alibaba. (Source: Crunchbase)

Total Funding: $607 million (Source: Crunchbase) And last but most certainly not least……

FaceX

FaceX was founded in 2018 in Bangalore, India. With a small team and limited amounts of investment, FaceX has been able to replicate and even improve on the results achieved by the top leaders in the Facial Recognition Industry. (As our tests will show)

Total Funding: $100,000

You may be wondering why we have mentioned the investments received by each of these companies. Investments play a huge role in how limited or unlimited a company’s resources can be and what is possible with a good investment.

Testing

For our tests, we picked two API features that are common across all 4 companies’ offerings. These are FACE DETECTION (with gender identification) and FACE VERIFICATION.

Face Detection – The ability of a facial recognition system to pick out faces in a static or moving image.

Face Verification – Analyzing two different pictures of a person and being able to identify whether or not both photos are of the same person.

Scoring: Face Detection – 10 point scale

             Gender Identification – 10 point scale

             Face Verification – 10 point scale

FINAL SCORE – OUT OF 30

NOTE: For both tests, stock photos from a Google image search were used. These tests were conducted on 14th March 2019 and are unbiased.

 

Face Detection (with Gender Identification)

For our face detection test, we used a stock image of a group of 30-40 people of different ethnicities and genders on a white background.

facial recognition industry leaders | Face Detection (with Gender Identification)

Face Verification

For the face verification test, we used 2 images of Captain America actor Chris Evans. The first image is from the first time, we saw him on screen as the superhero and the second image is from the most recent time he has appeared.

 

Face Verification

 

 

Face Verification test

 

 

 

The Results

NOTE: FOR THE GENDER IDENTIFICATION TEST, ALL 4 COMPETITORS MISGENDERED ONE PERSON IN PARTICULAR AND THEREFORE THE MAXIMUM ACHIEVABLE SCORE WILL BE 9/10.

 

Kairos

kairos

Kairos did really well in the face detection test. Being able to identify all faces in the image despite some heads being tilted. In the gender identification test, however, it performed poorly. It had wrongly identified the genders of 4 individuals in this image.

Face Detection: 10/10

Gender indentification: 6/10

Kairos result

 

 

In the face verification test, Kairos received the lowest score out of our contestants. It found no match between the two pictures and had a confidence of 63%.

Face Verification: 4/10

KAIROS FINAL SCORE: 20/30

  

Microsoft Azure Face API

Microsoft Azure Face API

Microsoft’s Face API put on the poorest performance for the face detection test, not being able to find more than half of the faces in the image. As for the gender test, we can only score it for the faces it identified and therefore receives a perfect score.

Face Detection: 2/10

Gender Identification: 9/10

microsoft azure face api result

For the face verification, Microsoft’s Face API was not able to match both images but received a better score since it had lower confidence of 38%.

Face Verification: 7/10

FINAL SCORE: 18/30

  

Face++

face++

Face++ performed decently in the face detection test. Unfortunately, it was not able to identify 2 faces in the image. In the gender identification test, it performed excellently. It had only misgendered the same person that all 4 platforms did.

Face detection: 7/10

Gender identification: 9/10

 

face++ results 

 

In the face verification test, Face++ performed the best being the only platform to correctly match the two faces.

Face Verification: 10/10

FINAL SCORE: 26/30

 

FaceX

facexThe best performing platform in both the face detection and gender identification test was FaceX. It had detected all the faces in the image and had identified all genders correctly except for the one.

Face Detection: 10/10

Gender Identification: 9/10

facex results

All things considered, FaceX’s face verification demo performed well. Despite not being able to match the faces, it had only a 13% confidence that the faces belonged to different people.

Face Verification: 9/10

FINAL SCORE: 28/30

 

THE VERDICT

FaceX’s platform outperformed the others overall with Face++ coming in second. Considering the size of both of these companies and the investments made into each of them, it is quite impressive what we have been able to achieve. At FaceX, we never stop learning. We improve our platform day by day to provide better services to our customers. When put on a level platform, we believe we can be among the top Faial recognition Industry Leaders as well and it is only a matter of time.

Leave a Comment

Your email address will not be published. Required fields are marked *