Efforts to identify bodies recovered after catastrophic flooding in Nepal have slowed to a crawl, hampered by a shortage of DNA specialists and laboratory capacity and by the condition of remains swept away in muddy floodwaters. Nepalese authorities are now using artificial intelligence facial recognition alongside DNA testing to match missing persons with recovered bodies.
A total of 1,377 bodies had been recovered as of the previous day, more than two weeks after the flooding struck the upper reaches of the Himalayas along the Nepal-China border on the 26th of last month, according to Nepalese outlet The Kathmandu Post and Spanish news agency EFE on the 11th. Of those, only 102 bodies — 7.4% of the total — had been identified and returned to families.
The main bottleneck is a lack of personnel and technical capacity for DNA testing. Only two facilities in Nepal can currently conduct such tests: the police Central Forensic Science Laboratory and the National Forensic Science Laboratory under the Ministry of Science, Technology and Innovation.
The two institutions normally process fewer than 1,000 samples a year. Since the flooding, more than 2,600 samples have arrived at once. Jiban Rijal, a DNA expert and former head of the forensic science laboratory, said both labs face a severe shortage of qualified specialists.
"We have always been short of staff," Rijal said. "The chemicals and reagents needed for testing are expensive and difficult to procure."
Nineteen DNA specialists from India and four from China have been deployed to the police Central Forensic Science Laboratory, and the Korea Disaster Relief Team has also joined the effort. Even so, identification work remains slow under an unprecedented volume of samples. Rijal said many of the missing are foreign nationals and stressed that authorities should draw on as many overseas experts as possible to identify the bodies quickly.
Damaged Remains Complicate DNA Collection, Prompting Shift to AI
Another obstacle is the state of the bodies themselves, many of which were swept away by powerful muddy currents or submerged for extended periods before being found. For badly damaged remains, Nepalese authorities are extracting DNA samples from teeth and bones, a process that takes longer than standard sampling.
With DNA testing alone proving too slow, AI has been brought in. NowtrKit, a U.S. artificial intelligence company, is providing an AI facial recognition system called FaceTag to support the identification work.
FaceTag analyzes faces captured in photographs and video and compares them with images of missing persons, narrowing the field to five to 10 possible matches.
Samar Panthi, the founder of NowtrKit, told Nepalese outlet Kantipur that about 500 missing persons have been registered with FaceTag so far and that 90 of them have been potentially identified.
Panthi said the company uses the technology to identify likely matches and then hands that information to Nepal Police or relevant government agencies for official verification.
AI cannot identify every body, however. Abhi Narayan Kaphle, a spokesman for Nepal Police, said authorities are working with technology companies to identify the dead but added that many of the bodies are too damaged for AI to identify.
