Spectacles that can 'read' text and convert it into speech to help the blind and visually impaired have been unveiled by scientists.

A small camera attached to the side of a pair of glasses uses OCR - optical character reading - technology to scan printed text and convert it into spoken words.

The camera is activated when the user points it at an object and presses a button, with the audio output fed into the earpiece of the person wearing the glasses.

Researchers at the University of California say this will help those with limited sight 'read' books, menus, newspapers and emails. And the camera is not limited to text: its makers say it can also recognise faces and products.


Lead researcher Dr Mark Mannis, whose work has been published in the journal JAMA Ophthalmology, said: 'Age-related macular degeneration is one of the most common causes of blindness in the elderly and it has no cure in its advanced stages.

'This device offers hope to patients who are beyond medical or surgical therapy for the condition.' He added: 'It is easily used and could potentially bring greater independence, particularly for older patients who are struggling with vision loss.'

His co-author Elad Moisseiev said: 'Patients with low vision are often dependent on handheld magnifiers, which may be cumbersome to use. But this is the first independent clinical study to evaluate this new low-vision-aid device based on novel optical-character recognition technology.'

He went on to say: 'Our results show that it can be a very useful aid for patients with low vision in performing activities of daily living, and increasing their functional independence.'

The devices are called OrCams and are on sale in the US for between US$2,500 and US$3,500.

Earlier this year it was tested by Luke Hines, 27, of Ilfracombe, Devon, who is blind in one eye and has only 3 per cent vision in the other after an operation to remove a childhood brain tumour.

He has described the device as 'life-changing' and said it means he will now be able to study at university.