We can make an app that blocks inappropriate Twitter content
Amsterdam, The Netherlands (PRWEB) March 24, 2014
Today, a new photo app for Android is released: Impala. The app recognizes scenes as food or architecture as soon as the camera is pointed. Scene recognition is realized by artificial intelligence running on the user’s phone, even before the picture is actually taken.
Based on what the smart photo camera "sees," photographic filters are automatically applied. The selected filters are optimized for popular scenes commonly photographed. For instance, when pointing the camera at people, a filter drawing anti-wrinkle hydrating masks over the faces of your friends is applied.
Furthermore, as soon as the camera detects you are about to take a picture of a handful of blueberries or a cinnamon scone, for example, it automatically applies a filter best fit for enhancing food colors. While taking photos of buildings or monuments, a filter best suited for outdoor scenes is applied. The altered lighting gives the images an edgier look.
“Computer vision and machine learning applications like these demand compute power and memory,” says Andrew Bagdanov, research fellow at the Universita Autonoma de Barcelona, Spain. “It is a challenge to get such applications to work on today’s mobile devices.”
A private beta version of the app was presented at Mobile World Congress (MWC) in Barcelona last month. "Conventions like MWC are massive technology overloads," says Edgar Cervantes, journalist at Phandroid / Neverstill Media. "It’s easy to miss smaller things when you have devices like the Samsung Galaxy S5 being announced, but this Impala app is quite the hidden treasure."
The Impala app uses artificial intelligence to recognize the contents of images and videos. Besides food, persons and architecture, the software recognizes sunsets, beaches, mountains, and so on. In addition to the camera functionality, the app also has a sorting functionality. For most smartphone users, the photo section is somewhat of a mess. Once the Impala app is installed, however, it automatically starts sorting all images on the phone into themed folders.
To achieve the real time performance, Euvision Technologies is one of the first companies in the world to utilize the Qualcomm Multicore Asynchronous Runtime Environment (MARE) software library. Qualcomm® MARE is a new developer programming library and API being developed by Qualcomm Research, a unit of Qualcomm Technologies, Inc., to enable Android native code developers to harness the benefits of multicore CPUs on modern smartphones & tablets.
“The new Impala for Android application highlights how quickly and easily mobile developers can now harness the multicore performance of Qualcomm® Snapdragon™ processors when developing in native code and using Qualcomm MARE," says Samir Kumar, director of Business Development and Product Management, Qualcomm Technologies, Inc. “Near real time and completely on device visual scene classification opens the door for a slew of rich contextual mobile imaging experiences that enables developers to create applications, helping make the Digital 6th Sense a reality.”
That said, another advantage of the technology is that it can prevent recording or viewing of unwanted scenes, such as classified military objects, on a mobile phone. As a technology demonstrator, the artificial intelligence component of the app has been trained to recognize hands. In the options section of the app, recording or displaying pictures of hands can be blocked.
"We can make an app that blocks inappropriate Twitter content,” Jan Willem Klerkx, CCO of Euvision Technologies, pointed out.
The app is created by the company Euvision Technologies. The firm plans to license its new mobile image classification engine to app developers and telecom operators. Today, the existing server version of the engine is used for moderation by social media and dating platforms, as well as for image searches by forensic agencies.
By putting Impala out there on the Google Play Store, the hope is to now introduce the technology to even more potentially licensing customers.
Qualcomm Snapdragon and Qualcomm MARE are products of Qualcomm Technologies, Inc. Qualcomm and Snapdragon are trademarks of Qualcomm Incorporated, registered in the United States and other countries, used with permission. Other product and brand names may be trademarks or registered trademarks of their respective owners.
About Euvision Technologies
The Impala app was developed by Euvision Technologies. The company is a spin-off from the University of Amsterdam and specializes in automated photo moderation and forensic searches. Euvision Technologies was founded in 2010 and employs 8 PhDs.