Link
Photo by Clive Darra
This post was originally posted on Jetpac CTO Pete Warden’s blog, June 9th.
I’m very pleased to announce that I’ve managed to port the Deep Belief image recognition SDK to the Raspberry Pi! I’m excited about this because it shows that even tiny, cheap devices are capable of performing sophisticated computer vision tasks. I’ve talked a lot about how object detection is going to be commoditized and ubiquitous, but this is a tangible example of what I mean, and I’ve already had folks digging into some interesting applications; detecting endangered wildlife, traffic analysis, satellites, even intelligent toys.
I can process a frame in around three seconds, largely thanks to heavy use of the embedded GPU for heavy lifting on the math side. I had to spend quite a lot of time writing custom assembler programs for the Pi’s 12 parallel ‘QPU’ processors, but I’m grateful I could get access at that low a level. Broadcom only released the technical specs for their graphics chip in the last few months, and it’s taken a community effort to turn that into a usable set of examples and compilers. I ended up heavily patching one of the available assemblers to support more instructions, and created a set of helper macros for programming the DMA controller, so I’ve released those all as open source. I wish more manufacturers would follow Broadcom’s lead and give us access to their GPUs at the assembler level, there’s a lot of power in those chips but it’s so hard to tune algorithms to make use of them without being able to see how they work.
Download the library, give it a try, and let me know about projects you use it on. I’m looking forward to hearing about what you come up with!
#deep learning#Pete Warden#Spotter#object recognition#object detection#tech#open source#technology#SDK#raspberry pi#raspberrypi
7 notes
·
View notes
Photo
If you'd been a computer you'd maybe look at the hotel as a palace and the pool as a giant tub? Try uploading a photo to see what spotter think it is!
0 notes
Photo
Spotting Restaurants With Beer
Our Spotter demo in javascript running directly in the browser could easily identify beers and restaurant from this photo. That's an example of how we are using the deep belief technology in our Jetpac City Guides app. In addition to spotting that this place is a restaurant and they're serving beer, we are also running the photo through an algorithm that can tell us if this photo is taken outdoors and if this place is frequented by people who tend to enjoy their drinks outdoors at restaurants with patios or at beer gardens. As a result, in the Jetpac City Guides app, you can find places popular with Outdoor Drinkers.
0 notes
Photo
#SpotterWin - Spaghetti Carbonara
#spagetthi#carbonara#spotterwin#Deep learning#deep belief#iOS#javascript#broswer#machine learning#Image processing#Object Recognition
0 notes
Photo

Spotter detected my dad's first camera from the 70s. #spotterwin #sx70 #deeplearning #deepbelief #polaroid #camera #objectrecognition #machinelearning #instanerd via Instagram http://ift.tt/1iwExW1
1 note
·
View note
Photo

Tennis ball spotted! #spotterwin #deeplearning #deepbelief via Instagram http://ift.tt/SiDN0b
0 notes
Photo
Yeah, kind of...
0 notes
Text
#SpotterFail
Playing with #deepbelief #deeplearning recognition @GA_DC pic.twitter.com/2ExWtlxAak
— Laura Lorenz (@LauraNLorenz) April 29, 2014
0 notes
Text
#SpotterWin - Coffee mug
Check out @JetpacApp Object Recognition http://t.co/FkoHeIflD3 #SpotterWin or #SpotterFail pic.twitter.com/WtJ2C03HlR
— ゆーきさん@ボロ雑巾 (@yuki_B) May 20, 2014
0 notes
Text
#Spotterfail - Espresso
Check out @JetpacApp Object Recognition http://t.co/XdieeofOB4 #SpotterFail pic.twitter.com/DtGp561SGc
— Andrew Barr (@APB_art) May 26, 2014
0 notes
Photo
We love the band name generator idea! claytonwoolard:
OTOH, it makes for an awesome band name generator.

App: Jetpac Spotter. “Tells you what things are when you point the camera at them.” Needs work.
124 notes
·
View notes
Photo
#SpotterWin
We were actually pretty impressed by Spotter getting this right!
#Spotterwin#Spotter#Jetpac#groom#javascript#Deep learning#deep belief#machine learning#Object Recognition
0 notes
Text
#SpotterWin - Home Theater
Cool new object identification app: @SpotterByJetpac. Check it out! Here's a #SpotterWin: pic.twitter.com/Xoz1LUHHNc
— Mikey Collard (@MikeyCollard)
April 24, 2014
0 notes
Text
#SpotterWin - Moped
Check out @JetpacApp Object Recognition http://t.co/duzGqFK8pI #SpotterWin or #SpotterFail pic.twitter.com/e2zZJiin2T
— Jade Glaze (@jadeglaze)
May 19, 2014
0 notes
Text
King Penguins VS Dogs Dressed Up As King Penguins
Experiment #1: Can Spotter tell the difference between a King Penguin and a dog dressed up as a King Penguin? To find out we tried the demo of Spotter working in the browser.
1. King Penguin
2. Dog dressed up as King Penguin
3. King Penguin
4. Dog dressed up as King Penguin
5. King Penguin
Conclusion: Spotter can tell a real King Penguin from a dog dressed up as a King Penguin.
0 notes