Breakthrough: China unveils AI eyes for the blind to move independently and safely
According to a study published in Nature Machine Intelligence, researchers from China unveiled a new wearable AI system that empowers blind and visually impaired individuals to navigate easily and independently. The system provides real-time guidance to users with a combination of video, vibrations, and audio prompts.
The AI system includes a camera, an AI processor, and bone conduction headphones. The system is mounted between the user's eyebrows. The camera captures the live footage instantly for the AI system to analyze. Then, short audio cues are delivered directly through the headphones, without blocking ambient sounds.
A team from Shanghai Jiao Tong University developed the system. Shanghai Artificial Intelligence Laboratory, East China Normal University, Hong Kong University of Science and Technology, and the State Key Laboratory of Medical Neurobiology at Fudan University.
'This research paves the way for user-friendly visual assistance systems, offering alternative avenues to enhance the quality of life for people with visual impairment,' the team wrote.
Lead researcher Gu Leilei, an associate professor at Shanghai Jiao Tong University, emphasised the team's commitment to making the system as practical and easy to use as possible.
"This system can partially replace the eyes,' he said.
'Lengthy audio descriptions of the environment can overwhelm and tire users, making them reluctant to use such systems,' Gu told The South China Morning Post.
'Unlike a car navigation system with detailed directions, our work aims to minimise AI system output, communicating information key for navigation in a way that the brain can easily absorb,' Gu said.
The equipment is lightweight and compact in design, meaning it can be worn all day without any discomfort, allowing users to move naturally without feeling burdened.
The system has been tested indoors with 20 visually impaired volunteers. After just 20 minutes of practice, most of the users could operate it with ease, according to the study.
The process to set a destination is also quite simple. Users issue a voice command, and the AI finds a safe route, offering only essential prompts along the way. It's currently trained to identify 21 objects, including beds, chairs, tables, sinks, televisions, and food items.
Researchers are planning to expand these capabilities further. Gu said the team's next focus is to refine the system for outdoor environments, where navigational challenges are far more complex. Enhancements could include improved object detection, dynamic route adaptation, and integration with real-world GPS systems.
This AI-powered wearable system may offer a new level of autonomy with further developments, and with that, confidence to millions of visually impaired people to be independent worldwide.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
8 hours ago
- Yahoo
Crypto Investors Lost $2.5B to Hacks and Scams in the First Half of 2025: Certik
Hackers stole more than $2.47 billion worth of cryptocurrency in the first half of 2025, already exceeding last year's total of $2.42 billion, according to Certik's Hack3d Report. The majority of that figure can be attributed to two incidents; the Bybit breach and the Cetus Protocol exploit that combined were worth $1.78 billion. Wallet compromise was the key attack vectors for hackers in the first half of this year, contributing to $1.7 billion worth of losses, while phishing was still rampant resulting in $410 million stolen across 132 incidents. Phishing is a technique hackers use to steal a victim's password or credentials in order to access an account. 'While the overall figures are alarming, it is important to point out that the majority of the funds lost in H1 were attributable to two concentrated, high-impact events,' said CertiK co-founder Ronghui Gu. 'But regardless, the results serve as another reminder to the industry that there is still much work to be done,' Gu added. The report also found that $801 million was lost in Q2 alone, a 52% decrease from the previous quarter. The majority of hacks took place on Ethereum, with $1.5 billion stolen across 164 incidents, followed by Bitcoin with $373 million stolen across 10 incidents. Sign in to access your portfolio
Yahoo
6 days ago
- Yahoo
Scientists Edited Sperm to Bypass an Egg. The Result? A Mouse With Two Dads.
"Hearst Magazines and Yahoo may earn commission or revenue on some items through these links." Here's what you'll learn when you read this story: The first mice to be created from two fathers, a phenomenon known as androgenesis, have now produced healthy offspring for the first time. Researchers implanted two sperm in an egg with its nucleus removed and modified the sperm with epigenome editing, turning certain genes on and off. This process carries a high risk, but it could be used for human reproduction and saving endangered species in the future. Millions of years of evolution have determined how mammals reproduce, but if something doesn't occur in nature, it is not necessarily impossible. It could still be achieved in a lab. Mammalian reproduction usually needs one sperm and one oocyte (egg) to create a zygote that develops in the womb until it becomes a new organism. Embryogenesis and fetal development in mice and humans are similar, which is why mice are often used as prototypes for human experiments. Also like humans, mice, at least on their own, cannot create a viable zygote out of two eggs or two sperm. That is where genetic intervention comes in. Earlier this year, researcher Yanchang Wei at Shanghai Jiao Tong University in China succeeded at creating mouse embryos from two sperm cells that were implanted into an egg after its nucleus had been removed. Giving the embryo a chance at survival meant making alterations to seven sites in the sperm DNA. The embryos grew inside surrogate mothers, but only two of over two hundred made it to adulthood. Those two mice have now reproduced successfully, showing that it is possible for at least one kind of mammal with two biological fathers to have healthy offspring of its own. 'Mammals are diploid creatures,' Wei said in a study recently published in PNAS. 'Every cell in their body contains two sets of chromosomes, one from the mother and one from the father. Mouse embryos with only paternal genomes can not develop to term, even being diploid (with two sets of chromosomes).' Behind what appears to be a miracle is a method known as epigenome editing. Because the epigenome is made up of chemical compounds that give the genome directions by modifying it, the process involves editing proteins, DNA and RNA at specific locations in the genome. There is no actual change to the genome made in epigenome editing, just a shift in which genes receive chemical signals to activate or inactivate them. There are some genes in sperm and eggs that are automatically switched on or off. Wei and his team used modified CRISPR proteins to turn on genes in the sperm cells that eggs would have normally had turned on. Androgenesis is rare among animals. While its does occur among some species of invertebrates, amphibians and fish. This phenomenon can happen when females produce eggs without a nucleus, which are then fertilized by males. The embryos will only inherit the genes of their father. In another type of androgenesis, both a sperm and egg form a zygote, but the genome of the female is eliminated. Males essentially use the eggs to clone themselves. It is thought that (with the exception of hermaphroditic organisms such as clams) androgenesis did not evolve much because the population would be dominated by male genes to the point of extinction. Creating mice from two mothers is easier in comparison. Known as parthenogenesis or 'virgin birth,' this phenomenon has been seen in reptiles, amphibians, insects, fish and even California condors. In 2004, a Japanese research team succeeded in genetically modifying and birthing the first fertile parthenogenic mouse, Kaguya. Whether genetically modified androgenesis or parthenogenesis will succeed in humans remains to be seen. The high risk is currently prohibitive. In the future, it might be a viable option, and could also be used to save critically endangered species which may only have a few individuals of one sex remaining. Never say 'impossible.' You Might Also Like The Do's and Don'ts of Using Painter's Tape The Best Portable BBQ Grills for Cooking Anywhere Can a Smart Watch Prolong Your Life?
New York Post
16-06-2025
- New York Post
AI is getting smarter every day —with thought processes already eerily similar to humans: study
Their powers go beyond AI-mpersonation. Artificial intelligence doesn't just look and act human — it supposedly thinks like us as well. Chinese researchers found the first-ever evidence that AI models like ChatGPT process information similarly to the human mind, detailing the dystopian-seeming discovery in the journal 'Nature Machine Intelligence.' Advertisement 'This provides compelling evidence that the object representations in LLMs (large language models), although not identical to human ones, share fundamental similarities that reflect key aspects of human conceptual knowledge,' wrote the team behind the study, which was a collaboration between the Chinese Academy of Sciences and the South China University of Technology, the Independent reported. 3 The AI models displayed remarkable powers of categorization. Malambo C/ – The team reportedly wanted to see if LLM models can 'develop humanlike object representations from linguistic and multimodal data (data in different forms such as text, audio, etc).' Advertisement 3 Prior to the study, many tech experts simply assumed that language models like ChatGPT mimicked human responses through pattern recognition. AlexPhotoStock – To discover whether the AI 'bot process' mirrors our cognition, researchers had OpenAI's ChatGPT-3.5 and Google's Gemini Pro Vision perform a series of 'odd-one-out' trials, in which they were given three items and tasked with selecting the one that doesn't fit, the South China Morning Post reported. 3 Researchers remain unsure if AI understands the significance or emotional value of cats and other entities on a human level. AntonKhrupinArt – Remarkably, the AI created 66 conceptual dimensions to sort the objects. Advertisement After comparing this cybernetic object sorting to human analysis of the same objects, they found striking similarities between the models' 'perception' and human cognition — particularly when it came to language grouping. From this, researchers deduced that our psychological doppelgangers 'develop human-like conceptual representations of objects.' 'Further analysis showed strong alignment between model embeddings and neural activity patterns' in the region of the brain associated with memory and scene recognition. Advertisement Researchers noted that the language-based LLMs were a bit lacking with regard to categorizing visual aspects such as shape or spatial properties. Meanwhile, research has shown that AI struggles with tasks that require deeper levels of human cognition, such as analogical thinking — drawing comparisons between different things to conclude — while it's unclear if they comprehend certain objects' significance or emotional value. 'Current AI can distinguish between cat and dog pictures, but the essential difference between this 'recognition' and human 'understanding' of cats and dogs remains to be revealed,' said He Huiguang, a professor at the Chinese Academy of Sciences' (CAS) Institute of Automation. Nonetheless, the scientists hope that these findings will allow them to develop 'more human-like artificial cognitive systems' that can collaborate better with their flesh-and-blood brethren.
