Menu
Frame-based Analysis for Your Videos
Frame-based Analysis for Your Videos

Implementation Considerations

Supported File Types

Currently, the Frame-based Analysis for Your Videos solution supports MP4 videos for frame processing and JPG images for facial searching. All face search images must use a .jpg extension. The solution will not process JPG images with a .jpeg extension.

Amazon S3 Server-Side Encryption (SSE)

AWS highly recommends that customers encrypt sensitive data in transit and at rest. This solution creates bucket policies on the video and face search buckets requiring objects to be encrypted at rest with Amazon S3 SSE. As a result, SSE must be enabled when uploading videos and images, such as using the --sse AES256 option with the AWS CLI. The following is a complete example for uploading a video using the AWS CLI:

aws s3 cp video_file.mp4 s3://$video_bucket/ --metadata topic=iot-topic-name --sse AES256

AWS IoT Topic

Each stage of the image search and video processing workflows are designed to provide status messages to an AWS IoT topic. This is an optional feature that you can use to monitor the video processing status for your videos.

To use this feature, create an x-amz-meta-topic header with a value of the desired AWS IoT topic name for status messages when you upload an image or video for processing. This header could contain a different topic name for each video, for each user, or for batches of videos to be processed, depending on the desired amount of video processing status granularity. This header can be configured manually through the Amazon S3 console or when uploading files with the AWS CLI with the --metadata topic=iot-topic-name option.

aws s3 cp video_file.mp4 s3://$video_bucket/ --metadata topic=iot-topic-name --sse AES256

Limits

This solution automatically uses managed Amazon Rekognition collections. It stores 100,000 face identifiers per collection. Once this number is reached, a new collection will be automatically created to index subsequent facial frames. When an image file is submitted to the face search feature, the face search Lambda function will search any Amazon Rekognition collections in parallel. If the face matches a facial frame in the collection, the percentage of the match will be shown in the svbp results table. The match threshold is currently set at 85%. If you want to increase or decrease the number of face identifiers per collection, you can modify the Max Collection Size parameter in the solution template.

Currently, this solution has been configured to support up to two parallel FFmpeg EC2 instances. While the Max Video parameter is configurable in the solution’s AWS CloudFormation template, we do not recommend changing this parameter. However, if you choose to change this parameter, you must request an increase to your AWS Lambda or Amazon Rekognition API limits and configure your Amazon DynamoDB capacity to meet your parallel video processing requirements.

Regional Deployments

This solution uses the Amazon Rekognition service which is currently available in specific AWS Regions only. Therefore, you must launch this solution in an AWS Region where Amazon Rekognition is available. For the most current Amazon Rekognition availability by region, see the AWS service offerings by region.