Perceptual Labs brings real time image and object detection to mobile and embedded devices.

 

 

Real Time Object Detection On Live Video on iPhone 

 

Why Real Time Object Detection?

security

Artificial Intelligence on device means data can't get compromised enabling applications where privacy is critical.  

bandwidth

Running Deep Learing  algorithms without the need to upload to a cloud server eliminates the need for costly bandwidth and can run in situations without connectivity.

speed

Taking advantage of the power of GPUs enables real time analysis on video feeds at 30 frames per second. Analyze and make decisions in real time.

any device

The ability to run on any mobile or embedded device at the edge is key to deploying solutions for real world applications.

Enabling applications in markets previously not accessible.

 

Medical

Enable remote diagnostic applications helping medical professionals diagnose more accurately, faster and where diagnostic expertise may not exist.

Internet of Things (IoT)

Adding AI to the edge creates smart sensors enabling a whole new class of IoT devices capable of performing actions in the field especially for bandwidth sensitive applications.

Agriculture

Using mobile devices in the field to diagnose crop diseases, weeds and nutrient deficiencies enable farmers to save a tremendous amount of money. Autonomous equipment and sensor aggregation will usher in the smart farm of tomorrow.

Industrial / Robotics

Creating smart robots where humans and robots can work together to perform complex tasks in a safe manner will lead to new factory concepts and efficiencies.

AR.jpg

Augmented Reality

Using object detection on device enables maintenance, training and consumer applications in real time and remote.

Intelligence

Analyze vast amounts of data to identify and classify specific assets in real time. Make critical decisions when time and accuracy is essential.

 

 

 

Tools, Workflow and Knowledge To Enable Your Application

 

Perceptual Labs has developed a framework to use optimized convolutional neural networks on many device types. The framework features an OpenGL-based rendering engine that performs GPU-accelerated convolutional network inference on a wide range of platforms, along with hardware-specific optimizations. It includes the ability to incorporate optimized network designs and efficiently run on device. The tools and workflow associated with taking an idea to release is what Perceptual Labs is about.

Examples From Live Video

airplane_iphone.png
dogbottle_iphone.png
personbicycle_iphone.PNG