- The Central Research Institute announced that after months of research, it has launched the world’s fastest facial recognition system with an accuracy rate higher than most others.
- The system is currently being used to detect objects such as cars and trains to regulate the traffic in Taiwan’s smart city in real-time.
- The researchers said that apart from being used for traffic regulation, the technology can be used to detect those not wearing a mask and calculate those flouting social distancing rules.
Facial recognition technology is far from a new concept. Tech giants in the likes of Facebook, Google, Microsoft, Amazon, and Apple are in the mix and have launched their own programs by leveraging AI using it for various purposes.
Indeed, the use of facial recognition has exploded in a slew of sectors. Lately, through a new deal, Taiwan and Russia have announced the development of the world’s fastest facial recognition technology.
The Central Research Institute said that the object detection algorithm deemed YOLOv4 is currently the fastest image recognition technology and has been trialed in the traffic system of Taiwan’s smart city Taoyuan and Hsinchu.
Jointly developed by Taiwan scholars Wang Jianyao and Liao Hongyuan along with Russian developer Alexey Bochkovskiy, the technology has reported an accuracy rate of 43.5% which is a 10% increase since its previous version. YOLOv4 is an abbreviation of ‘You only look once,’ the firm said.
Detailing on YOLOv4, the Chinese Academy of Sciences highlighted that the algorithm leverages AI for real-time detection of objects and faces. It combines image sensors and computer vision to detect objects such as train speed at every intersection in real-time. The three researchers have been upgrading the previous versions since November last year to make the system faster and remove any potential glitches.
“It has the advantages of light system architecture and efficient algorithms. It can be calculated using a general graphics processor (GPU), ‘like using an abacus the speed of the computer,’ the firm added.
Commenting on the trial in Taiwan, Hongyuan said that the smart traffic analysis system powered by the YOLOv4 technology, which has been deployed in Taoyuan and Hsinchu, can instantly detect vehicles in real-time, stop trains and slow its speed at every intersection therefore increasing safety.
Jianyao added that cities using this modern technology will experience more efficient road conditions and older traditional systems will become obsolete. He said:
“Our current system is [installed] at every intersection. It transfers data as soon as it receives and finishes the analysis, so it can directly affect people in a traffic jam and can improve the current situation.”
Academia Sinica highlighted that technology has a host of use cases apart from smart traffic. For instance, in the current times dominated by global pandemic Covid-19, nations can detect people who do not wear masks and calculate whether people maintain a social distance to better control the spread.
The tech can also be used to regulate traffic flow, ease factory defect detection, medical image analysis as well as boost the self-driving car sector for further research and development.
Facial recognition sparks controversy despite uses due to biases
The technology might be one of the most divisive concepts. While the AI-based tech has its advantages, it can be misused in a way that it becomes detrimental to certain groups of people. In the US, cops used facial recognition to identify protesters by linking their pictures to social media profiles leading to police misconduct with citizens of colour.
“IBM firmly opposes and will not condone uses of any [facial recognition] technology, including facial recognition technology offered by other vendors, for mass surveillance, racial profiling, violations of basic human rights and freedoms, or any purpose which is not consistent with our values and Principles of Trust and Transparency,” CEO Arvind Krishna said in a letter to the Congress. “We believe now is the time to begin a national dialogue on whether and how facial recognition technology should be employed by domestic law enforcement agencies.”
In addition, in a blog post published on 10 June, Amazon said it will suspend the use of its technology by police for a year therefore giving time for federal laws protecting human rights to be imposed.
While bias and invasion of privacy are risks that AI entails, time will tell whether these issues can be resolved completely and if law enforcement will use it responsibly.