Jump to Main ContentJump to Primary Navigation
Top

Pigeons can spot cancer with the accuracy of a human

bird.jpg

What a coup for science.

A team of trained pigeons has successfully demonstrated it can spot a cancerous tumour in mammogram images - with the same accuracy as a human.

That's not to suggest the birds pooled their talents and flocked over to the nearest lab to show off their skills: a team led by Richard Levenson of the University of California and Edward Wasserman at the University of Iowa trained 16 birds to "peck" a button when they are presented with a computer image of a cancerous mammogram scan.

Contrary to their idiotic behaviour - not moving out of the way of traffic, walking like muppets, crapping everywhere - pigeons actually have a visual advantage over humans for this task: they view the world with five different colours, as opposed to our three, and their brains don't "fill in" gaps or create patterns like ours do (which is why we're constantly seeing faces in pieces of burnt toast or clouds).

The study team trained the pigeons by placing them in a box with a computer screen. Images of previously diagnosed breast cancer histopathology slides were shown to the birds, along with a blue and yellow button. If the pigeon correctly identified a cancer cell, they were rewarded with a food pellet. If they were wrong, they received nothing. 

No, the pigeons didn't know what they were looking at - it's a form of pattern recognition. The birds learn in a matter of hours to distinguish cancerous from noncancerous cells, with an accuracy of 80 per cent after a month of training. While a single bird wasn't as capable as a single human, when the birds worked as a flock (the same samples shown to 16 birds) they returned an accuracy of 99 per cent - putting them on par with humans.

"The birds’ successes and difficulties suggest that pigeons are well-suited to help us better understand human medical image perception", wrote the research team, "and may also prove useful in performance assessment and development of medical imaging hardware, image processing, and image analysis tools."

While it's not proposed that pigeons replace lab technicians just yet, flocks of cell spotters could free up human researchers to carry out more complicated tasks. Like making tea. 

[Via: Science Mag]

Related

rexfeatures_1930076a.jpg

Your terrible drunken behaviour might not actually be your fault

tinder.jpg

7 Easy ways to get more Tinder matches (using science)

sleep5.png

This is what would happen if you didn’t sleep

space.jpg

Anyone can understand space travel with this excellent cartoon

tumblr_lziif0S8HI1rp784so1_500.jpg

Great white sharks are massive fans of metal music

coffeesuperhero1.jpg

You should probably be drinking more coffee (says science)

Comments

More

Leaked poster appears to confirm Destiny 2 release date

And it sounds like they're doing a beta in June

by Matt Tate
23 Mar 2017

Android users can finally play Super Mario Run - and definitely should

Extended toilet breaks for everyone

by Matt Tate
23 Mar 2017

Apple have released a red iPhone and that is very important

And not just because it's red

by Gary Ogden
21 Mar 2017

Amazon's new Alexa update means it can bring you beer in two hours

"Alexa, we're going to need more booze"

by Matt Tate
21 Mar 2017

Forget traffic jams with this SUV that can drive over cars

Why has it taken so long to invent this?

by Dave Fawbert
21 Mar 2017

Zelda megafan controls his smart home with an ocarina

Pointless? Possibly. Are we envious? Definitely

by Matt Tate
20 Mar 2017

Soon you'll be able to steal your mate's phone battery to charge your

"Oh come on, mate, I've only got 4%"

by Gary Ogden
17 Mar 2017

How to cheat your way to victory in 'Mario Kart 64'

Anyone fancy digging the old N64 out again?

by Matt Tate
17 Mar 2017

5 new(ish) mobile games guaranteed to make any commute bearable

It doesn't have to be this painful

by Matt Tate
16 Mar 2017

This terrifying gadget allows you to make private calls in public

Be like Bane

by Gary Ogden
15 Mar 2017