Digital video surveillance is the fastest growing market now, valued at US$3bil (RM10.3bil) this year, according to Tang Juel Hoi, CEO of Digital Info Tech (DIT) Pte Ltd.
This is because technological advances in recent years have made digital video surveillance more feasible than the traditional videotape-based system, he said.
For instance, huge amounts of video data can now be transmitted in real-time via wider Internet bandwidth and faster connection speeds. "In the past, storage was always a problem because video files are usually very large," Tang said.
Some systems try to apply compression techniques to shrink file sizes, but usually at the cost of picture quality.
"Sometimes the police would ask us to help reconstruct a criminal's face captured on video because the quality is so poor," said Tang. "But now with the digital video system, that is less of a problem.
"I can imagine that things would be even better when 3D holographic storage becomes available next year." (See P18.)
Until then, video compression is still important vis-a-vis storage. DIT is developing a "next-generation" video compression process that it claims would be better than the current H.264 advanced video-coding standard.
Tang was speaking at the recent DIT Global Security Perspectives Forum in Petaling Jaya. DIT, a subsidiary of the Singapore-based Thakral Corporation Ltd, is a provider of digital lifestyle products and technology solutions.
More uses
Another reason why digital video surveillance is more popular nowadays, Tang said, is because people have found more uses for it.
"You may be able to capture everything that happens at a place through video surveillance, but most of this data will be junk save for a few important details and events," he said.
This is where intelligent processing comes into the picture. The system can single out details like human faces through pattern-matching techniques, according to him.
For example, whenever an intelligent system detects a human face, it can put a box around it to ease the monitoring process for security guards.
Furthermore, intelligent processing can be coupled with human behaviour analysis to create smarter monitoring systems.
Using motion detection techniques, a system can be programmed to tell whether a person is walking, running or falling, said Tang.
In Singapore, there are swimming pools which are equipped with such a system. Some swimming pools even have a system that can highlight swimmers in different colours according to how dangerously they swim.
"Those swimming normally are in green, while those who are drowning are highlighted in red," said Tang.
For the masses
Digital video surveillance can also be used together with biometric systems to create better security systems, said Teh Tuan Ann, DIT production innovation director.
He said that many governments around the world are looking at implementing biometric-based identification systems to improve national security, especially in the areas of fingerprint and facial recognition.
"Compared to other biometric-based systems, face recognition is the most natural because that's basically what humans use to identify each other," Tang said.
However, a face-recognition system has its weaknesses – generally it has poor accuracy, is slow and difficult to use.
"We are partnering with the Chinese Academy of Sciences and Singapore's Agency for Science, Technology and Research to develop better systems using predictive face synthesis," Teh said.
Next month, DIT will be launching its Flexcam Face Recogniser, a face recognition-based authentication software that provides enhanced security when logging into a PC.
Flexcam identifies a user's face via an attached webcam and permits or denies access accordingly.
During the user enrolment or registration process, the software will capture five samples of the user's face, either automatically or manually.
It will take note of the user's facial features and generate a 3D model of the entire face using predictive techniques.
"This is a step up from the old method of storing only 2D static photographs of a user in a database," said Teh.
"The next time the same user logs in, Flexcam will detect his face and verify it within a second or two," he said during a demonstration of the system.
Targeted at consumers, Flexcam will cost about US$30 to US$40 (RM102.60 to RM136.80).
