Chinese scientists have made a breakthrough by developing optical fibre technology that could enable real-time endoscopies of brain nerve cells and accelerate next-generation telecommunications.
The newly developed “super-thin fibre,” operating on an optical neural network, could carry tens of thousands of times more optical information than traditional single-mode fibers.
Researchers from the University of Shanghai for Science and Technology, Southeast University in Nanjing, and the University of Technology Sydney published their findings in the peer-reviewed journal Nature Photonics on Friday.
Optical fibres, or fibre-optic cables, are thin strands of plastic or glass that transmit data as pulses of light at high speed.
While there are existing solutions to reconstruct scrambled images, including using artificial neural networks or spatial light modulators, the process is time- and energy-consuming, with computational delays.
It involves converting light signals into electrical signals, which are then processed and interpreted by an AI model.
Also read: Not for sale': OpenAI turns down Musk’s $97.4 billion takeover bid.
Yu Haoyi, an associate professor at the University of Shanghai for Science and Technology’s School of Artificial Intelligence Science and Technology, stated that the team is working with a regional hospital on Hainan Island in southern China to test the super-thin fibre for use in minimally invasive endoscopies.
“We’ve initiated the use of fibre-integrated diffractive neural networks to detect light signals indicating lesions in samples obtained from the hospital under a microscope,” he said.
This marks the first step toward testing the technique on animals, followed by humans.
“Discussions with doctors have guided the direction of our research,” he added.
“Patients could ingest harmless luminescent materials that make targeted organs glow, enabling our optical-fibre device to detect any abnormalities within the body.”
Yu further explained that the new technology achieves higher image resolution than current endoscopic equipment, allowing it to detect developing or premature tumours that may not be visible due to image distortion.
“We can also integrate the device with a deep-learning algorithm trained to identify pathological changes, improving screening accuracy and facilitating early detection, diagnosis, and treatment,” he said.
Also read: Musk makes $97.4 billion bid to acquire OpenAI