New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[jsk_perception/aws_auto_checkin_app] Use search_faces_by_image() of Amazon Rekognition #2716
Conversation
…TimeSynchronizer option
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let me clearfy the contents of this PR
- fix typo and color convert, support approxymate_sync/queue_size in approxymate sync: thanks for fix!
- support use_aws_detection/use_usb_cam/gui args in
sample_aws_auto_checkin_app.launch
: This is also great improvement. Do you have any comment on comparison between aws_face_detection and OpenCV's face detection, for example performance and speed? Please add that to document https://github.com/jsk-ros-pkg/jsk_recognition/blob/master/doc/jsk_perception/nodes/aws_auto_checkin_app.rst#sample
I think I wrongly useduse_window
args to bothsample_aws_auto_checkin.launch
andsample_aws_detect_faces.launch
, may be you can remove that and usegui
. Please also adddoc="..."
togui
. - Using draw_rects.py to publish result images: good visualization is always welcome.
I think above fixes can be merged immediately
--
Could you explan more about using recognition / dynamodb ?
a) I can not reproduce botocore.exceptions.NoCredentialsError: Unable to locate credentials
error. Because of using old libraries?
$ pip freeze| grep boto
boto3==1.4.2
botocore==1.16.19
b) I am not familier with dynamodb, but how is it differnet from current auto_cehckin_apps? Do they have better performance then current checkin? or do they have different interfaces? for example, how to add new faces? currently we put the face image to auto-check-in-gapp-register... buckets. Or it is just a creanup API of old aws checkkin example?
Basically libraries alwyas update, and our time is limited. So we should follow new API only when we really need this and we concentrate on improving your own/unique idea.
c) why you need new aws.json
file? if I need to do something on existing aws.json file, please let me know, we'd like to use same aws.json for all aws_ examples.
I fixed it by 9a10b5d.
aws_face_detection (1 ~ 50 Hz) seems to be slow compared with OpenCV's face detection (30 ~ 500 Hz), but the result is more accurate. OpenCV often incorrectly recognizes faces.
I am using the same libraries (boto3==1.4.2 and botocore==1.16.19) and followed the instruction written in jsk doc,
Since I cannot execute current checkin and aws_detect_faces has no error, I applied boto3 client to aws_auto_checkin in the same way.
This PR needs the change of Anyway, the condition of this problem should be checked first. |
I can reproduce the error in docker environment (18.04).
|
See https://gist.github.com/k-okada/b8f4fac48f8e473c4ae9328b916a0176 for the output of |
self.roi_sub = message_filters.Subscriber('face_roi', FaceArrayStamped) | ||
self.ts = message_filters.ApproximateTimeSynchronizer([self.image_sub, self.roi_sub], 10, 1, allow_headerless = True) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please consider using
self.transport_hint = rospy.get_param(
namespace + 'image_transport', 'compressed')
so that we can switch compressed image / raw image
c.f.
http://wiki.ros.org/image_transport/Tutorials/ExaminingImagePublisherSubscriber
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you. raw image becomes available by 0fdd1d8
Thank you for the result of Also, I created a simple test program. However, the result was nocredentialserror.
@k-okada |
sorry, I had my own I also ask to change "env.json" as follows.
|
I confirmed that the old Note that, the new
Could you change the |
…t on use_aws_face_detection
I changed |
hi, i have added AmazonDynamoDBReadONlyAccess polcy to rekognition users, so can we use old aws.json file? If so, I think we we want to keep using old files. |
@k-okada |
Thanks. I deactivated my |
Without this change,
client.initiate_auth
has failed withbotocore.exceptions.NoCredentialsError: Unable to locate credentials
.This error is solved by using AWS SDK for Python, boto3.
Following changes accompanying:
cc: @iory