Developer Challenge Showcases Potential of Computer Vision

AWS DeepLens offers cameras and pre-built ML models to help developers gain hands-on experience with image recognition and more.

Computer vision is generating excitement in applications like security surveillance and traffic management, but now there's a chance for developers to help everyday people with innovative and imaginative twists on this new tech.

The AWS DeepLens Challenge covers four categories -- inclusivity, sustainability, games, and health -- that will have a positive impact on the world. Developers everywhere are invited to leverage the AWS DeepLens camera and AWS machine learning services to create something good and support related nonprofit organizations, as well. That's right: Every entrant gets $50 in AWS credits, and the charity associated with each category gets a $249 donation.


At the first AWS DeepLens Challenge back in May, developers came up with several cool projects, including:

  • ReadToMe, which uses AWS DeepLens and Amazon Polly to recognize text and read it aloud.
  • Dee, an interactive game that asks children to answer questions by picking corresponding images.
  • SafeHaven, an app for people living alone that uses AWS DeepLens together with Amazon Rekognition to build a database of trusted faces and issue an alert when strangers are at the door.

Machine learning on AWS DeepLens

When AWS announced the release of AWS DeepLens this summer, developers were excited by the opportunity to experiment with computer vision and machine learning hands-on.

The AWS DeepLens camera includes a 4-megapixel sensor, 2D microphone array, WiFi connectivity, and an Intel Atom processor, enabling the camera to run algorithms locally instead of sending images to the cloud for processing.

The package also comes with a comprehensive toolset, including pre-built ML models for object and activity recognition. Developers can start tinkering with computer vision in as little as 10 minutes, according to AWS, then customize their own models on Amazon SageMaker in the cloud before deploying them to the camera.

Four ways to help

For the inclusivity challenge, AWS is encouraging developers to think about any situation where individuals may be excluded due to disability, race, gender, or even geographic location or education level. In one early use case, an Australia-based developer named Chris Coombs trained a AWS DeepLens camera to recognize and translate the American Sign Language alphabet. The associated charity is Northwest Center, an organization that helps children and adults with developmental disabilities.


How can computer vision help include everyone?

Click the image to share your ideas.

For the sustainability challenge, any AWS DeepLens project that promotes eco-friendly activities and respect for the environment will qualify. In one example, AWS engineers created an ML model that enables AWS DeepLens cameras to recognize different containers and determine whether they are recyclable. In another, developer Paul Langdon created Backyard Birder, a AWS DeepLens model that can recognize various species of birds and keep a tally of birds and squirrels that visit a backyard feeder. Engineers for a Sustainable World -- a group that boasts over 1,750 members working on solutions to global problems in 13 countries around the world -- is the charity that benefits from this challenge category.


How can computer vision help the environment?

Click the image to share your ideas.

In the games challenge, AWS is looking for ways to create or improve gaming using AWS DeepLens. Developers are encouraged to consider sports, puzzles, and multiplayer or single-player games. One cool example is Simon Says, a computer vision-aided take on the classic game where AWS DeepLens can determine whether players did the correct action and whether they did it when Simon said. Girls Who Code -- which offers free after-school programs and career guidance for girls in middle school and high school who are interested in technology -- is the beneficiary of this challenge.


How would you add computer vision to games?

Click the image and rise to the challenge.

For the health challenge, developers can draw inspiration from projects such as DermLens, in which a AWS DeepLens camera can recognize and track the progress of psoriasis treatments, and Exercise Counter, where the camera tracks the sets and reps in an exercise regimen and creates statistics over time. Hopecam, an organization that arranges for video equipment so that immuno-compromised children can stay connected to their classes in school, will benefit from entries in this category.


Want to promote health with a smart camera?

Click the image and bring your idea to life.

You have from now until January 31, 2019 to submit your projects. Choose your challenge and get started today! Register now for the AWS DeepLens Challenge!