An extract from another interesting article from the team at "Havard Business Review". Click on the link to find our more or register for their regular updates: https://augmented-startups.storychief.io/
Everything to Know About the Face Detection Technology
Facial recognition technology was a part of the science fiction genre. This technology not only became a reality but also is widespread. It is difficult to read technology news without seeing something posted about face detection. Different industries are benefiting from this technology including law enforcement agencies, retailers, airports, and smartphone companies.
When Did Facial Technology Come Into Existence? - The Brief History
Facial technology has been under development for many decades. We will share a brief history of how this technology came into existence.
1. Manual Measurement (1960s)
Woodrow Wilson Bledsoe develops the RAND tablet system, which could manually record the coordinate locations of facial features including eyes, nose, mouth, and hairline.
2. The 21 Markers (1970s)
Goldstein, Harmon, and Lesk add accuracy to the manual facial recognition system in the 1970s with 21 specific markers to identify faces.
3. EigenFaces (1980-1990)
Sirovich and Kirby apply linear algebra to facial recognition. In 1991, Turk and Pentland expanded EigenFaces system with detecting faces using images.
4. FERET Program (1993-2000)
The DARPA (Defense Advanced Research Projects Agency) and National Institute of Standards (NIS) and Technology provide the Face detection Technology (FERET) in the 1990s.
5. Face detection Vendor Testing (2000s)
The National Institute of Standards and Technology (NIST) begins to conduct Face detection Vendor Tests during the early 2000s.
6. Installation of Face detection Technology in an Airport (2011)
The government of Panama works with the US Secretary of Homeland Security for the FaceFirst face detection platform to stop illegal activities in the Tocumen airport in Panama.
7. Watchlist as a Service (2017)
FaceFirst introduces WatchList as a Service (WaaS) at the NRF Protect conference. The WaaS is a new face detection platform to prevent shoplifting and other violent crime.
What Are the Recent Advancements in the Field of Face Detection Technology?
The facial recognition technology has been in development since the 1960s. Recent technological advancements have provided it with endless capabilities. It is no longer limited to only science fiction movies, but millions of viewers now have access to this technology. We have compiled a list of activities, which face detection technology would help accomplish in the future.
1. Online Payments
Businesses are looking for an easy way to get payments from the consumer. The online shopping has shown how smooth a cashless transaction can be. The introduction of face detection technologies as FaceTech would make it easier for customers to shop online even without their credit or debit cards. The MasterCard also launched a new application named MasterCard Identity Check to help consumers pay by using their phone cameras.
Medical professionals would be able to recognize any illnesses of a person by looking at their features. It would alleviate the ongoing rush in medical centers by reducing waiting lists and simplifying the entire process. Patients need to decide as to whether it is worth waiting a whole month for an appointment or undergo a cyber scan to help the doctors detect any issue. Biometrics can also help the medical experts store a patient’s data by using their photograph instead of a username and password.
3. Criminal Identification
The United States Federal Bureau of Investigation is working on using machine learning algorithm to identify suspects using their driver’s license. At present, the FBI has a big database of suspects including half of the population faces. In short, the face detection technology will help the government take care of the criminal elements in the society.
What is FaceNet?
The FaceNet is a Deep Learning architecture, which includes convolutional layers based on the GoogLeNet. It returns a 128-dimensional vector embedding for each face. It trained with triplet loss for different classes of faces to capture similarity and differences between human faces. It is an advanced face detection technology.
The 128-dimensional embedding returned by the FaceNet model clusters faces. The vector is closer for similar faces and distant for different faces. The architecture trained over a dataset with a large number of faces belonging to different classes. One can teach the SVM classified or other multi-class classifiers over the vector embedding obtained for faces. Every time, a new person’s face added to the dataset, one would need to add another class and train the classifier.
Top 5 Facial Recognition Platforms
Many internet companies including Facebook and Google have included face detection into their artificial intelligence systems. Facebook has started training both face and image recognition models depending on hashtags.
Many smartphone companies have introduced facial recognition technology in their devices. Samsung launched its Galaxy Note 7 with Iris recognition. The technology of facial recognition worked wonders for sectors using biometric identification systems including healthcare, travel, banking, and retail. Given below are the major face detection technologies to expect this year or are dominant in the market:
1. Google Cloud Vision
The combination of facial recognition and cloud engines is unbeatable. If a user application runs on Google cloud engine, it will become easier to integrate the Cloud Vision into the product or application. AutoML can train customer vision models. The Google Cloud Vision can detect multiple faces within an image. Google is presently working on adding a facial recognition feature to it.
2. Microsoft Face API
The Microsoft Face API can identify similar faces and organize the face images into related groups. It lets the application developers tag people in pictures, which include security and internet applications. Face API provides Face Verification as a service, which one can use in case of two different faces of the same person. The Face Detection service can detect one or more human faces in an image. Other features of this face detection technology include emotion recognition such as anger, contempt, disgust, fear, happiness, and other emotions.
Kairos specializes in facial recognition and detection. It offers facial recognition solutions, and customers prefer to host Kairo on their servers, as it provides data security. The face analysis features include face detection, identification, and verification.
4. IBM Watson Visual Recognition
IBM Watson Visual Recognition widely utilized in computer vision applications. It can help programmers identify the name and type of person with an image. At present, this face detection technology supports formats as .GIF, .jpg, .png, and .tif.
5. Amazon Rekognition
The Amazon Rekognition is another giant leap in the field of facial recognition technology. It can quickly identify activism, people, and faces among any given image content. The most popular feature of Rekognition is the facial analysis service, which offers a relatively accurate facial recognition engine working on different images and videos.
What is the Future of Facial Recognition Technology?
The facial recognition technology has been around since the 1960s, but recent technological advancement has provided it endless capabilities. It is no longer limited to only science fiction movies, but millions of viewers now have access to this technology. We have compiled a list of activities, which face recognition technology would help accomplish in the future.
Prevent shoplifting and other crimes
Find missing persons
Help law enforcement agencies
Protect Public places from threats
Control Access to Sensitive Areas
Future of face detection would help detect and identify faces in complex backgrounds.
The face detection technology has a lot more to offer to the public than mere surveillance. It is the one thing, which can help in detecting not only a person but also the criminal elements.