Posted On: Apr 8, 2019
We are excited to announce that AWS DeepLens has a new project template for Bird Classification which enables classification of 200 species of birds.
This project yields the top 5 predictions on bird species from a static bird photo captured by the AWS DeepLens camera. The model uses ResNet-18 neural network architecture, and has been trained using CUB-200 dataset to identify 200 different bird species. To learn more, visit the AWS DeepLens documentation. To dive deep into another method on how you can implement 'Bird Classification' by modifying Amazon SageMaker’s built-in Object Detection algorithm to classify birds using AWS DeepLens, read this blog.
It's easy to get started with AWS DeepLens using the sample projects provided which cover some of the most popular computer vision use cases. These include Object Detection, Artistic Style Transfer, Face Detection, Hot Dog Not Hot Dog, Cat and Dog Recognition, Action Recognition, Head Pose Detection, and the new Bird Classification. Each project template consists of a default machine learning model and AWS Lambda function to trigger the execution of the model.
As your skills and ideas develop, you can build custom deep learning models in the cloud using Amazon SageMaker. Check out the collection of projects created with AWS DeepLens by the developer community for inspiration.