With Kenya getting ready to install a Kshs 45.3 billion Integrated Public Safety Communication and Surveillance System (IPSCSS) in a bid to fight the rising cases of insecurity, experts have come out to say more needs to be done to make the system effective.

Kenya’s leading telco, Safaricom, was recently given the green light by a parliamentary committee Administration and National Security to install over 1,800 ultra-high definition CCTV cameras said to have unique facial recognition capabilities, with those familiar to the system saying that it could pull out a criminal’s face from a crowd.

ITWeb Africa sought to find out just how practical this is, and went ahead to speak to Eliud Mwaniki, currently a security consultant having worked in the past with various security firms including G4S Singapore.

Mwaniki said that installation of the IPSCSS will be the first step in the right direction, however quickly adding that the security situation in the country should not be expected to change overnight.

“I think the integrated National Surveillance, Communication and Control System will help -at least in the long run- to deal with the rising cases of criminal activities in the country, ranging from petty crimes like pickpockets to sophisticated terrorist activities,” Mwaniki told ITWeb Africa.

“However, with the system relying on facial recognition software to identify and track criminals, Kenya will need to do a lot more in ensuring the system’s capabilities are explored to the maximum,” Mwaniki said.

He said that in countries where the system was already being used, there exists well-managed system updating processes, which ensure the surveillance systems receive a steady upgrading and updating with more recent pictures of the citizens and anyone entering their territories.

“For a surveillance system, specifically one that relies on images to identify faces, to be effective, the computer requires several training images so as to ensure the special features on a face, like the space between the eyes, the shapes of the jaw bones, relative positions of the cheekbones and other biometrics,” Mwaniki told ITWeb Africa.

“However, most of the photos in the government’s custody are those that were captured during issuance of national identity cards and passports, images that are not sufficient to train the surveillance system in facial recognition,” Mwaniki continued.

He said that there is a lot that needs to be done including updating the images, something that could pose a significant challenge to those working on the system.

“If this is not done, the system could be reduced to just one that will be used to monitor activities in public places, with people sitting behind the screens observing these activities. This will require a lot of human power, and there are possibilities some crimes will go unnoticed to the observer’s eyes, based on the number and size of the screens,” Mwaniki said.

Mwaniki further explained that another unique challenge facing Kenya is that the country does not have a well-organised address system to ensure criminals are tracked to their doorstep once the facial recognition identifies them.

Mwaniki also said that an overhaul of the country’s databases is needed, to guarantee the system works as effectively as it does in other countries.

“There is no doubt that the system is well advanced, and can be a part of the solution in this rising insecurity in the country. But unless other government departments and operations are integrated upgraded and constantly updated, then we might not see the full advantages of installing the multi-billion system,” Mwaniki concluded.