Create a Heatmap with Amazon Rekognition

This project adds a heatmap layer on top of a picture based on Amazon Rekognition detect labels function.

Here is an example picture from Amazon Bring-Your-Kids-To-Work Day. We can add a HeatMap layer on top of it and see which game attracked more people.

Setup

Prerequest

1. Create an S3 bucket.
2. Upload the picture that you want to add a heatmap layer into the S3 bucket you just created.

Run the demo

This demo runs the code using Amazon SageMaker. You can choose any IDE you like to run the python code.

1. Create a Notebook instance in Amazon SageMaker.
   a. Notebook instance type: t2.medium.
   b. IAM role: Create a new role.
2. Add Rekognition access permission to SageMaker.
   a. Click the instance you just created.
   b. In the Permissions and encryption section, click the link under IAM role ARN. This will bring you to the IAM page.
   c. Click the Attach policies button. Find the AmazonRekognitionFullAccess policy and attach it.
   d. Then go back to the SageMaker console page. Click Open Jupyter button.
3. Download the HeatMap_Code.ipynb file from here and upload it into your Jupyter notebook.
4. Open the HeatMap_Code in the Jupyter notebook and change the required parameters (bucketName and imageName) at the beginning.
5. Then you can run the code and a HeatMap layer will be added on your picture.

Customize your HeatMap

You can customize your HeatMap by changing the parameters in the optional section at the beginning when open the Jupyter notebook.

1. matrixhight: This parameter control the number of rows you will get on the HeatMap grid. The default number is 20.
2. matrixwidth: This parameter control the number of columns you will get on the HeatMap grid. The default number is 30.
3. DarkestColor: The HeatMap will provide gradient colors between the darkest color you set and the color white. The default color is Crimson Red (#990000). You can find more color code here.
4. LabelType: This HeatMap layer was first created to detect human group. The default value for this parameter is "Person". You can change this to any other labels that Amazon Rekognition support such as cars, trees, etc. It can be used to detect the density of other objects as well.

Clean Up

1. Go to Amazon SageMaker console, choose Notebook Instances.
2. Select the instance you created for this demo. Click Actions and choose Stop from the drop down list.
3. Once the instance status becoming Stopped, select this instance and click the Actions again. Choose Delete from the drop down list to delete this instance.
4. Go to Amazon S3 console, delete the bucket you created for this demo.

Security

See CONTRIBUTING for more information.

License

This library is licensed under the MIT-0 License. See the LICENSE file.