PR.com: Business Directory, Press Releases, Jobs, Products, Services, Articles
 
Businesses Articles Press Releases Follow @PRcom
Press Release Pricing | News by Category | News by Country | News by US Region | Recent News | PR.com News on Your Site
 

Press Releases

 
ZZ Wolf

Press Release


Receive press releases from ZZ Wolf: By Email RSS Feeds:

ZZ Photo 2.2 Utilizes Deep Neural Networks for Pet Detection


Ukrainian startup ZZ Photo announces version 2.2 of its photo organizer, which is now equipped with deep neural networks to automatically search for cats and dogs in photographs. The new technology is called "DeepPet" and allows detection of pets in photos 4-5 times more accurately than in previous versions.

Kiev, Ukraine, September 26, 2015 --(PR.com)-- ZZ Photo application users who already benefitted from the product´s many photo organization features, will now be able to do so much faster and more accurately. The main difference in the new version of the ZZ Photo organizer is the implementation of Deep Learning algorithms for image recognition.

Having emerged in the late 2000s, Deep Learning is a revolutionary artificial intelligence technology that allows the training of artificial neural networks on large volumes of data. Many of the most innovative and advanced product features currently available, such as facial recognition on Facebook and Google, as well as Apple's Siri's voice recognition, are all based on Deep Learning.

The new "DeepPet" pet detecting technology in ZZ Photo provides 4-5 times more accurate detection of cats and dogs in photos compared to traditional object recognition algorithms based on Viola-Jones cascades of classifiers and other similar methods. Deep convolutional neural networks are used as the main building blocks of the new technology. Since these neural networks are large, their training must be done by machines equipped with powerful NVidia CUDA video cards to speed up the mathematical calculations. Neural networks inside ZZ Photo were trained for two-category pattern recognition problem, so they are working at least at 1,5-2 times more accurate than standard "out-of-the-box" deep neural networks trained for ImageNet, PASCAL VOC and similar problems with many image categories.

ZZ Photo team plans to develop its Deep Learning technology into a separate API http://recognizz.it, meant to be used in other commercial projects not limited to task of photo organizing.

Once installed on a user's computer, ZZ Photo can detect directories containing photos and allows the user convenient access to them, via both automatic and semi-automatic modes for sorting and arranging photos. Currently ZZ Photo is equipped with facial recognition and pet detection algorithms as well as the ability to identify scanned documents, search for duplicates and more. The application can also synchronize with the user´s Facebook and Instagram image albums.

Price and availability

ZZ Photo is available free of charge for Windows 7/8/8.1/10.

Links

Website: http://zzphoto.me
Video about ZZ Photo: https://www.youtube.com/watch?v=--Zh66nbAGw
Contact Information
ZZ Wolf
Julia Wunder
7 391 278 95 26
Contact
zzwolf.com

Click here to view the list of recent Press Releases from ZZ Wolf
Promote Your Business