Discover Your Island

Island University Researcher Receives $100,000 Grant from Amazon to Revolutionize Robotic Warehouse

March 28, 2017

Rahnemoonfar-Amazon-Webstory-450x299.jpg

CORPUS CHRISTI, Texas – You can teach an old dog tricks, but can you teach a program newer, better tricks? Dr. Maryam Rahnemoonfar, Assistant Professor of Computer Science at Texas A&M University–Corpus Christi, will be working with Amazon, the largest internet-based retailer, to do just that. She has recently been awarded $80,000 by the Amazon Academic Research Awards (AARA) program for her proposal “Real Time Heterogeneous Product Counting on Amazon Bin Image Data Set based on Deep Learning.” She also received additional funding of $20,000 in Amazon Web Services (AWS) Cloud Credits to pursue her research.

“I am excited to received such a prestigious award with the maximum allowed amount from Amazon, a leading industry in the high-tech field,” said Rahnemoonfar. “I am looking forward to closely work with Amazon to develop fundamental algorithms in deep learning and computer vision that will address the challenges they are currently facing.”

What is deep learning in artificial intelligence? Rahnemoonfar suggests thinking of it like the human brain, which also has a hierarchy of processing data.

“When humans recognize a face or a voice, although it seems simple and instantaneous, our brain is performing a deep and hierarchal processing between our senses and a set of neurons that we then link to that particular person,” explained Rahnemoonfar.

How does recognizing objects fit in with Amazon? Across the nation, there are various warehouses filled with bins of Amazon products, waiting for their buyers. Not too long ago, the Amazon Fulfillment Technologies team released the Amazon Bin Image Data Set which contains thousands of images of what inside these bins. Right now, these warehouses are human-run, but the goal is to include robotic assistance. That’s where Rahnemoonfar’s research comes in.

“The goal of my project is to create real time object detection and counting in the Amazon Bin Image Data Set,” said Rahnemoonfar. “Our new advances in computer vision and deep learning can revolutionize Amazon’s robotic warehouses.”

The research team will be using a simulation algorithm to create a data set that produces information and number of each product. They will also be designing a new deep learning architecture to help the program work outside of conditions in a lab environment. The research will take place over one year.

Rahnemoonfar also has high hopes that this project will lead to greater partnerships in the future.

“I hope this project will be an opening for a great collaboration of TAMU-CC with leading industries and that it brings several internships and employment opportunities for our students,” said Rahnemoonfar.

The research team consists of undergraduate and graduate students in Computer Science within the Bina Lab. The students are currently working on a variety of deep learning algorithms for applications including autonomous driving, aging in place, yield estimation and autonomous unmanned aerial vehicles.