Detecting text in a stored
video
Amazon Rekognition Video text detection in stored videos is an asynchronous operation. To start
detecting text, call StartTextDetection.
Amazon Rekognition Video publishes the completion status of the video analysis to an Amazon SNS topic. If the video analysis is
successful, call GetTextDetection
to get the analysis results. For
more information about starting video analysis and getting the results, see Calling Amazon Rekognition Video operations.
This procedure expands on the code in Analyzing a video stored in an Amazon S3
bucket with Java or Python (SDK). It uses an Amazon SQS queue to get the
completion status of a video analysis request.
To detect text in a video stored in an Amazon S3 bucket (SDK)
-
Perform the steps in Analyzing a video stored in an Amazon S3
bucket with Java or Python (SDK).
-
Add the following code to the class VideoDetect in step 1.
- Java
-
//Copyright 2018 Amazon.com, Inc. or its affiliates. All Rights Reserved.
//PDX-License-Identifier: MIT-0 (For details, see https://github.com/awsdocs/amazon-rekognition-developer-guide/blob/master/LICENSE-SAMPLECODE.)
private static void StartTextDetection(String bucket, String video) throws Exception{
NotificationChannel channel= new NotificationChannel()
.withSNSTopicArn(snsTopicArn)
.withRoleArn(roleArn);
StartTextDetectionRequest req = new StartTextDetectionRequest()
.withVideo(new Video()
.withS3Object(new S3Object()
.withBucket(bucket)
.withName(video)))
.withNotificationChannel(channel);
StartTextDetectionResult startTextDetectionResult = rek.startTextDetection(req);
startJobId=startTextDetectionResult.getJobId();
}
private static void GetTextDetectionResults() throws Exception{
int maxResults=10;
String paginationToken=null;
GetTextDetectionResult textDetectionResult=null;
do{
if (textDetectionResult !=null){
paginationToken = textDetectionResult.getNextToken();
}
textDetectionResult = rek.getTextDetection(new GetTextDetectionRequest()
.withJobId(startJobId)
.withNextToken(paginationToken)
.withMaxResults(maxResults));
VideoMetadata videoMetaData=textDetectionResult.getVideoMetadata();
System.out.println("Format: " + videoMetaData.getFormat());
System.out.println("Codec: " + videoMetaData.getCodec());
System.out.println("Duration: " + videoMetaData.getDurationMillis());
System.out.println("FrameRate: " + videoMetaData.getFrameRate());
//Show text, confidence values
List<TextDetectionResult> textDetections = textDetectionResult.getTextDetections();
for (TextDetectionResult text: textDetections) {
long seconds=text.getTimestamp()/1000;
System.out.println("Sec: " + Long.toString(seconds) + " ");
TextDetection detectedText=text.getTextDetection();
System.out.println("Text Detected: " + detectedText.getDetectedText());
System.out.println("Confidence: " + detectedText.getConfidence().toString());
System.out.println("Id : " + detectedText.getId());
System.out.println("Parent Id: " + detectedText.getParentId());
System.out.println("Bounding Box" + detectedText.getGeometry().getBoundingBox().toString());
System.out.println("Type: " + detectedText.getType());
System.out.println();
}
} while (textDetectionResult !=null && textDetectionResult.getNextToken() != null);
}
In the function main, replace the lines:
StartLabelDetection(bucket, video);
if (GetSQSMessageSuccess()==true)
GetLabelDetectionResults();
with:
StartTextDetection(bucket, video);
if (GetSQSMessageSuccess()==true)
GetTextDetectionResults();
- Java V2
-
This code is taken from the AWS Documentation SDK examples
GitHub repository. See the full example here.
//snippet-start:[rekognition.java2.recognize_video_text.import]
import software.amazon.awssdk.auth.credentials.ProfileCredentialsProvider;
import software.amazon.awssdk.regions.Region;
import software.amazon.awssdk.services.rekognition.RekognitionClient;
import software.amazon.awssdk.services.rekognition.model.S3Object;
import software.amazon.awssdk.services.rekognition.model.NotificationChannel;
import software.amazon.awssdk.services.rekognition.model.Video;
import software.amazon.awssdk.services.rekognition.model.StartTextDetectionRequest;
import software.amazon.awssdk.services.rekognition.model.StartTextDetectionResponse;
import software.amazon.awssdk.services.rekognition.model.RekognitionException;
import software.amazon.awssdk.services.rekognition.model.GetTextDetectionResponse;
import software.amazon.awssdk.services.rekognition.model.GetTextDetectionRequest;
import software.amazon.awssdk.services.rekognition.model.VideoMetadata;
import software.amazon.awssdk.services.rekognition.model.TextDetectionResult;
import java.util.List;
//snippet-end:[rekognition.java2.recognize_video_text.import]
/**
* Before running this Java V2 code example, set up your development environment, including your credentials.
*
* For more information, see the following documentation topic:
*
* https://docs.aws.amazon.com/sdk-for-java/latest/developer-guide/get-started.html
*/
public class DetectTextVideo {
private static String startJobId ="";
public static void main(String[] args) {
final String usage = "\n" +
"Usage: " +
" <bucket> <video> <topicArn> <roleArn>\n\n" +
"Where:\n" +
" bucket - The name of the bucket in which the video is located (for example, (for example, myBucket). \n\n"+
" video - The name of video (for example, people.mp4). \n\n" +
" topicArn - The ARN of the Amazon Simple Notification Service (Amazon SNS) topic. \n\n" +
" roleArn - The ARN of the AWS Identity and Access Management (IAM) role to use. \n\n" ;
if (args.length != 4) {
System.out.println(usage);
System.exit(1);
}
String bucket = args[0];
String video = args[1];
String topicArn = args[2];
String roleArn = args[3];
Region region = Region.US_EAST_1;
RekognitionClient rekClient = RekognitionClient.builder()
.region(region)
.credentialsProvider(ProfileCredentialsProvider.create("profile-name"))
.build();
NotificationChannel channel = NotificationChannel.builder()
.snsTopicArn(topicArn)
.roleArn(roleArn)
.build();
startTextLabels(rekClient, channel, bucket, video);
GetTextResults(rekClient);
System.out.println("This example is done!");
rekClient.close();
}
// snippet-start:[rekognition.java2.recognize_video_text.main]
public static void startTextLabels(RekognitionClient rekClient,
NotificationChannel channel,
String bucket,
String video) {
try {
S3Object s3Obj = S3Object.builder()
.bucket(bucket)
.name(video)
.build();
Video vidOb = Video.builder()
.s3Object(s3Obj)
.build();
StartTextDetectionRequest labelDetectionRequest = StartTextDetectionRequest.builder()
.jobTag("DetectingLabels")
.notificationChannel(channel)
.video(vidOb)
.build();
StartTextDetectionResponse labelDetectionResponse = rekClient.startTextDetection(labelDetectionRequest);
startJobId = labelDetectionResponse.jobId();
} catch (RekognitionException e) {
System.out.println(e.getMessage());
System.exit(1);
}
}
public static void GetTextResults(RekognitionClient rekClient) {
try {
String paginationToken=null;
GetTextDetectionResponse textDetectionResponse=null;
boolean finished = false;
String status;
int yy=0 ;
do{
if (textDetectionResponse !=null)
paginationToken = textDetectionResponse.nextToken();
GetTextDetectionRequest recognitionRequest = GetTextDetectionRequest.builder()
.jobId(startJobId)
.nextToken(paginationToken)
.maxResults(10)
.build();
// Wait until the job succeeds.
while (!finished) {
textDetectionResponse = rekClient.getTextDetection(recognitionRequest);
status = textDetectionResponse.jobStatusAsString();
if (status.compareTo("SUCCEEDED") == 0)
finished = true;
else {
System.out.println(yy + " status is: " + status);
Thread.sleep(1000);
}
yy++;
}
finished = false;
// Proceed when the job is done - otherwise VideoMetadata is null.
VideoMetadata videoMetaData=textDetectionResponse.videoMetadata();
System.out.println("Format: " + videoMetaData.format());
System.out.println("Codec: " + videoMetaData.codec());
System.out.println("Duration: " + videoMetaData.durationMillis());
System.out.println("FrameRate: " + videoMetaData.frameRate());
System.out.println("Job");
List<TextDetectionResult> labels= textDetectionResponse.textDetections();
for (TextDetectionResult detectedText: labels) {
System.out.println("Confidence: " + detectedText.textDetection().confidence().toString());
System.out.println("Id : " + detectedText.textDetection().id());
System.out.println("Parent Id: " + detectedText.textDetection().parentId());
System.out.println("Type: " + detectedText.textDetection().type());
System.out.println("Text: " + detectedText.textDetection().detectedText());
System.out.println();
}
} while (textDetectionResponse !=null && textDetectionResponse.nextToken() != null);
} catch(RekognitionException | InterruptedException e) {
System.out.println(e.getMessage());
System.exit(1);
}
}
// snippet-end:[rekognition.java2.recognize_video_text.main]
}
- Python
-
#Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
#PDX-License-Identifier: MIT-0 (For details, see https://github.com/awsdocs/amazon-rekognition-developer-guide/blob/master/LICENSE-SAMPLECODE.)
def StartTextDetection(self):
response=self.rek.start_text_detection(Video={'S3Object': {'Bucket': self.bucket, 'Name': self.video}},
NotificationChannel={'RoleArn': self.roleArn, 'SNSTopicArn': self.snsTopicArn})
self.startJobId=response['JobId']
print('Start Job Id: ' + self.startJobId)
def GetTextDetectionResults(self):
maxResults = 10
paginationToken = ''
finished = False
while finished == False:
response = self.rek.get_text_detection(JobId=self.startJobId,
MaxResults=maxResults,
NextToken=paginationToken)
print('Codec: ' + response['VideoMetadata']['Codec'])
print('Duration: ' + str(response['VideoMetadata']['DurationMillis']))
print('Format: ' + response['VideoMetadata']['Format'])
print('Frame rate: ' + str(response['VideoMetadata']['FrameRate']))
print()
for textDetection in response['TextDetections']:
text=textDetection['TextDetection']
print("Timestamp: " + str(textDetection['Timestamp']))
print(" Text Detected: " + text['DetectedText'])
print(" Confidence: " + str(text['Confidence']))
print (" Bounding box")
print (" Top: " + str(text['Geometry']['BoundingBox']['Top']))
print (" Left: " + str(text['Geometry']['BoundingBox']['Left']))
print (" Width: " + str(text['Geometry']['BoundingBox']['Width']))
print (" Height: " + str(text['Geometry']['BoundingBox']['Height']))
print (" Type: " + str(text['Type']) )
print()
if 'NextToken' in response:
paginationToken = response['NextToken']
else:
finished = True
In the function main, replace the lines:
analyzer.StartLabelDetection()
if analyzer.GetSQSMessageSuccess()==True:
analyzer.GetLabelDetectionResults()
with:
analyzer.StartTextDetection()
if analyzer.GetSQSMessageSuccess()==True:
analyzer.GetTextDetectionResults()
- CLI
-
Run the following AWS CLI command to start detecting text in a
video.
aws rekognition start-text-detection --video "{"S3Object":{"Bucket":"bucket-name","Name":"video-name"}}"\
--notification-channel "{"SNSTopicArn":"topic-arn","RoleArn":"role-arn"}" \
--region region-name --profile profile-name
Update the following values:
-
Change bucket-name and
video-name to the Amazon S3 bucket name and file
name that you specified in step 2.
-
Change region-name to the AWS region that
you're using.
-
Replace the value of profile-name with the
name of your developer profile.
-
Change topic-ARN to the ARN of the Amazon SNS
topic you created in step 3 of Configuring Amazon Rekognition Video.
-
Change role-ARN to the ARN of the IAM
service role you created in step 7 of Configuring Amazon Rekognition Video.
If you are accessing the CLI on a Windows device, use double
quotes instead of single quotes and escape the inner double quotes
by backslash (i.e. \) to address any parser errors you may
encounter. For an example, see below:
aws rekognition start-text-detection --video \
"{\"S3Object\":{\"Bucket\":\"bucket-name\",\"Name\":\"video-name\"}}" \
--notification-channel "{\"SNSTopicArn\":\"topic-arn\",\"RoleArn\":\"role-arn\"}" \
--region region-name --profile profile-name
After running the proceeding code example, copy down the returned
jobID and provide it to the following
GetTextDetection command below to get your results,
replacing job-id-number with the jobID you
previously received:
aws rekognition get-text-detection --job-id job-id-number --profile profile-name
-
Run the code. Text that was detected in the video is shown in a list.
Filters
Filters are optional request parameters that can be used when you call
StartTextDetection. Filtering by text region, size and confidence
score provides you with additional flexibility to control your text detection
output. By using regions of interest, you an easily limit text detection to the
regions that are relevant, for example, a bottom third region for graphics or a top
left corner for reading scoreboards in a soccer game. Word bounding box size filter
can be used to avoid small background text which may be noisy or irrelevant. And
lastly, word confidence filter enables you to remove results that may be unreliable
due to being blurry or smudged.
For information regarding filter values, see DetectTextFilters.
You can use the following filters:
-
MinConfidence –Sets the confidence
level of word detection. Words with detection confidence below this level
are excluded from the result. Values should be between 0 and 100.
-
MinBoundingBoxWidth – Sets the
minimum width of the word bounding box. Words with bounding boxes that are
smaller than this value are excluded from the result. The value is relative
to the video frame width.
-
MinBoundingBoxHeight – Sets the
minimum height of the word bounding box. Words with bounding box heights
less than this value are excluded from the result. The value is relative to
the video frame height.
-
RegionsOfInterest – Limits
detection to a specific region of the frame. The values are relative to the
frame dimensions. For objects only partially within the regions, the
response is undefined.
GetTextDetection response
GetTextDetection returns an array (TextDetectionResults)
that contains information about the text detected in the video. An array element,
TextDetection, exists
for each time a word or line is detected in the video. The array elements are sorted
by time (in milliseconds) since the start of the video.
The following is a partial JSON response from GetTextDetection. In
the response, note the following:
-
Text information – The
TextDetectionResult array element contains information
about the detected text (TextDetection)
and the time that the text was
detected in the video (Timestamp).
-
Paging information – The example
shows one page of text detection information. You can specify how many text
elements to return in the MaxResults input parameter for
GetTextDetection. If more results than
MaxResults exist, or there are more results than the
default maximum, GetTextDetection returns a token
(NextToken) that's used to get the next page of results.
For more information, see Getting Amazon Rekognition Video analysis results.
-
Video information – The response
includes information about the video format (VideoMetadata) in
each page of information that's returned by
GetTextDetection.
{
"JobStatus": "SUCCEEDED",
"VideoMetadata": {
"Codec": "h264",
"DurationMillis": 174441,
"Format": "QuickTime / MOV",
"FrameRate": 29.970029830932617,
"FrameHeight": 480,
"FrameWidth": 854
},
"TextDetections": [
{
"Timestamp": 967,
"TextDetection": {
"DetectedText": "Twinkle Twinkle Little Star",
"Type": "LINE",
"Id": 0,
"Confidence": 99.91780090332031,
"Geometry": {
"BoundingBox": {
"Width": 0.8337579369544983,
"Height": 0.08365312218666077,
"Left": 0.08313830941915512,
"Top": 0.4663468301296234
},
"Polygon": [
{
"X": 0.08313830941915512,
"Y": 0.4663468301296234
},
{
"X": 0.9168962240219116,
"Y": 0.4674469828605652
},
{
"X": 0.916861355304718,
"Y": 0.5511001348495483
},
{
"X": 0.08310343325138092,
"Y": 0.5499999523162842
}
]
}
}
},
{
"Timestamp": 967,
"TextDetection": {
"DetectedText": "Twinkle",
"Type": "WORD",
"Id": 1,
"ParentId": 0,
"Confidence": 99.98338317871094,
"Geometry": {
"BoundingBox": {
"Width": 0.2423887550830841,
"Height": 0.0833333358168602,
"Left": 0.08313817530870438,
"Top": 0.46666666865348816
},
"Polygon": [
{
"X": 0.08313817530870438,
"Y": 0.46666666865348816
},
{
"X": 0.3255269229412079,
"Y": 0.46666666865348816
},
{
"X": 0.3255269229412079,
"Y": 0.550000011920929
},
{
"X": 0.08313817530870438,
"Y": 0.550000011920929
}
]
}
}
},
{
"Timestamp": 967,
"TextDetection": {
"DetectedText": "Twinkle",
"Type": "WORD",
"Id": 2,
"ParentId": 0,
"Confidence": 99.982666015625,
"Geometry": {
"BoundingBox": {
"Width": 0.2423887550830841,
"Height": 0.08124999701976776,
"Left": 0.3454332649707794,
"Top": 0.46875
},
"Polygon": [
{
"X": 0.3454332649707794,
"Y": 0.46875
},
{
"X": 0.5878220200538635,
"Y": 0.46875
},
{
"X": 0.5878220200538635,
"Y": 0.550000011920929
},
{
"X": 0.3454332649707794,
"Y": 0.550000011920929
}
]
}
}
},
{
"Timestamp": 967,
"TextDetection": {
"DetectedText": "Little",
"Type": "WORD",
"Id": 3,
"ParentId": 0,
"Confidence": 99.8787612915039,
"Geometry": {
"BoundingBox": {
"Width": 0.16627635061740875,
"Height": 0.08124999701976776,
"Left": 0.6053864359855652,
"Top": 0.46875
},
"Polygon": [
{
"X": 0.6053864359855652,
"Y": 0.46875
},
{
"X": 0.7716627717018127,
"Y": 0.46875
},
{
"X": 0.7716627717018127,
"Y": 0.550000011920929
},
{
"X": 0.6053864359855652,
"Y": 0.550000011920929
}
]
}
}
},
{
"Timestamp": 967,
"TextDetection": {
"DetectedText": "Star",
"Type": "WORD",
"Id": 4,
"ParentId": 0,
"Confidence": 99.82640075683594,
"Geometry": {
"BoundingBox": {
"Width": 0.12997658550739288,
"Height": 0.08124999701976776,
"Left": 0.7868852615356445,
"Top": 0.46875
},
"Polygon": [
{
"X": 0.7868852615356445,
"Y": 0.46875
},
{
"X": 0.9168618321418762,
"Y": 0.46875
},
{
"X": 0.9168618321418762,
"Y": 0.550000011920929
},
{
"X": 0.7868852615356445,
"Y": 0.550000011920929
}
]
}
}
}
],
"NextToken": "NiHpGbZFnkM/S8kLcukMni15wb05iKtquu/Mwc+Qg1LVlMjjKNOD0Z0GusSPg7TONLe+OZ3P",
"TextModelVersion": "3.0"
}
