Facial recognition technology has subtly been integrated into our everyday applications over the last few years. For example, Facebook knows which friends to tag in our photos and our photo applications know how to group our photos by people. However, this technology is being taken further than social media and is being tested in a variety of industries for different purposes. Some of these applications include an increase in security by Customs and Border Protection, government and business facilities that require a high level of security clearance, and real-time criminal searches using closed-circuit televisions. Aside from these applications of facial recognition, it is most notably being used by large retailers. This is mostly being done in an effort to compete with large online shopping websites. Through the use of this technology, retailers will be able to instantly recognize customers and provide them targeted offers based on their buying preferences. When you first appear in their system, retailers begin to build up a profile based on your in-store actions. This includes the amount of time you spend in a certain area and an individual’s path around the store. This, in turn, allows marketers to know their customers better in an effective manner. Artem Kukharenko, CEO of a company developing facial recognition technology, Ntechlab, stated “Visitor’s photos may function as cookies referring to the identification and storage of user settings. In other words, loyalty cards will become obsolete. As soon as you walk into a store, the staff will already know what you bought last time, thanks to the camera’s footage and our technology.”
On one hand, this technology will prove to be really helpful for retailers who face great competition from online companies such as Amazon and other online stores that implement targeted offers. One of the issues with this type of marketing is that, unlike other means of personally identifiable biometric data collection, facial recognition works from a distance. This means that people do not have the option to opt out of this form of data collection if the cameras used in retail stores indiscriminately identify shoppers and store their actions performed while shopping. Another issue with facial recognition being used in this way is how this sensitive biometric data is stored. I believe that these companies should not have the right to store such information in their databases without the consent of the individual. Lastly, facial recognition systems are susceptible to false-positive identifications. This can surely lead to serving irrelevant advertisements to consumers. Facial recognition technology should be developed further and regulated properly before its ubiquitous adoption. Reference
9 Comments
Researchers at the University of California, Los Angeles, in collaboration with ObjectVideo, developed a prototype computer vision system that has the ability of generating a live text transcription of what is happening in a surveillance camera feed. Although the application of this system was only for surveillance footage, this could potentially be used to help with searching video in the future, since searching for images and video is done using the surrounding text. As input, this system takes an array of images or video frames and generates a summary of what what is going on in those images. These images can then be searched using plain text. Aside form video search, this system addresses the increase in surveillance cameras being used and the increased number of people needed to survey these cameras. The automation of surveillance camera surveying is accomplished in the system by first separating the objects in an image from their background. Then, the collection of the objects identified is used to extract meaning from the current video scene. The object meaning part of the system comes from human intelligence. Prior to the development of this system Song-Chun Zhu, lead researcher and professor of statistics and computer science at UCLA, created a nonprofit that helped in the creation of a database consisting of more than two million images containing objects that were classified into more than 500 categories. All of this aggregated data also gave the system the ability to describe the movement of objects in successive frames.
There is definitely great benefit in being able to analyze security video footage in real time, including being able to respond to a situation quicker and identifying persons of interest. However, there are also ways that people can abuse this system. A system like this can be used to indiscriminately target certain behavior, and falsely generalize someone for something they might have accidentally done. Also, this system can be used to specifically target an individual without their knowledge or prior consent and track their movements. I believe that a system like this must be carefully regulated in order to ensure that no harm is done to people knowingly or unknowingly. Reference |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |