Computer Vision

At PARC, we’re using novel approaches in deep learning and computer vision to help organizations gain valuable insights from images and solve complex, real-world problems.

What is Computer Vision?
Bigger Insights, Better Results
Why Work with PARC?
Applications

What is Computer Vision?

Computer vision is the science by which machines learn to understand and reason with the world through images and video. Not surprisingly, today’s most successful computer vision approaches employ deep neural networks—inspired by the human brain—to accomplish a breadth of complex visual tasks.

Computer vision allows machines to identify people, places, objects and activities in images and video with:

• Accuracy at or above human levels
• Greater speed and efficiency

Bigger Insights, Better Results

By extracting meaningful information from images and videos, computer vision can help organizations across industries achieve better results:

• Increase operational efficiency and reduce costs
• Enhance the customer experience
• Improve product quality
• Gain valuable insights that drive better business decisions

Why Work with PARC?

Our team has a broad range of expertise in computer vision, spanning both image and video analysis. Whether you’re looking to detect moving objects, analyze images on a smartphone, or develop a new augmented reality application, we can help you explore the possibilities.

Holistic Approach to Computer Vision
Systems Engineered for Accuracy and Efficiency

Our scientists optimize the interplay between image, hardware, software and human to build highly robust systems that perform complex visual tasks.

• Unique combination of hardware and software expertise:
  • Building deep visual learning algorithms with domain knowledge and models
  • Understanding and optimizing computer vision systems end- to-end for best performance
  • Balancing and optimizing edge- versus cloud-based visual computing
  • Drawing on our specialty in optics, sensors and cameras—including PARC’s unique, low-cost and compact hyperspectral camera
  • Working with a wide range of data inputs from diverse sources—including RGB, spectral, depth, and geolocation
• Bringing humans into the loop:
  • We understand how humans and machines augment each other in solving complex visual tasks
  • Our applications provide interactive feedback during image capture to ensure only quality images are used for analysis
  • We use augmented reality to interactively annotate objects for training computer vision systems
  • Our solutions automate routine tasks, freeing up humans to focus on more creative endeavors

Small Data, Big Insights
Get Better Results Without the Investment of Big Data

Deep learning boosts our capabilities in computer vision—allowing for object detection, image classification and more. However, traditional deep learning requires large quantities of annotated data, which can be difficult, time-consuming and costly for organizations looking to explore new applications.

Using novel techniques, we expand the benefits of deep learning to applications with limited datasets and reduce the time and cost attached to innovation.

Innovation with Commercial Impact
Optimizing the Customer Experience

We take the work inside of a lab and translate it into practical business solutions. Leveraging our expertise in computer vision, user experience design, and strategic business models, we focus on creating the most value for our partners and their customers.

Applications

The applications for computer vision are far reaching.  Here are just a few examples of our work:

• Health & Beauty
  Detect and analyze facial skin condition from smartphone images. Provide product recommendations to help consumers make informed choices
  Learn about our work with Olay
• Transportation
  Reduce maintenance costs and prevent unplanned downtime through asset monitoring and fault detection
  Learn about our work with East Japan Railway Company
• Workplace of the Future
  Provide augmented reality assistance for complex enterprise tasks, such as machine repair and maintenance
• Medical Diagnostics
  Accurately detect subtle patterns and anomalies in MRIs, CAT scans and X-rays
  Learn about our work with Penn State
• Retail
  Use video analytics to gain insights into store traffic and shopping patterns, and enhance the shopping experience
• Insurance
  Use images captured by drones and other aerial platforms to assess property condition and damage
• Personalized Marketing
  Improve customer engagement and retention through personalized image and video content creation
• Additive Manufacturing
  Apply image and video analysis to perform real-time monitoring of 3D printing process