issue with multi-person tracking with PoseNet using JavaScript & TensorFlow
We are using an issue with multi-person tracking with the PoseNet library for Javascript. We are seeing the data getting associated with the wrong person sometimes.
person 1
- tracking data
person 2
- tracking data
some time later:
person 1
- tracking data of person 2
person 2
- tracking data of person 1
We can add some code in Javascript to check the validity of the data. Is this the way multi-person tracking works in PoseNet or a bug?
javascript tensorflow
asked Nov 20 '18 at 22:01
M. BediM. Bedi
946815
add a comment |
We are using an issue with multi-person tracking with the PoseNet library for Javascript. We are seeing the data getting associated with the wrong person sometimes.
person 1
- tracking data
person 2
- tracking data
some time later:
person 1
- tracking data of person 2
person 2
- tracking data of person 1
We can add some code in Javascript to check the validity of the data. Is this the way multi-person tracking works in PoseNet or a bug?
javascript tensorflow
asked Nov 20 '18 at 22:01
M. BediM. Bedi
946815
add a comment |
We are using an issue with multi-person tracking with the PoseNet library for Javascript. We are seeing the data getting associated with the wrong person sometimes.
person 1
- tracking data
person 2
- tracking data
some time later:
person 1
- tracking data of person 2
person 2
- tracking data of person 1
We can add some code in Javascript to check the validity of the data. Is this the way multi-person tracking works in PoseNet or a bug?
javascript tensorflow
asked Nov 20 '18 at 22:01
M. BediM. Bedi
946815
We are using an issue with multi-person tracking with the PoseNet library for Javascript. We are seeing the data getting associated with the wrong person sometimes.
person 1
- tracking data
person 2
- tracking data
some time later:
person 1
- tracking data of person 2
person 2
- tracking data of person 1
We can add some code in Javascript to check the validity of the data. Is this the way multi-person tracking works in PoseNet or a bug?
javascript tensorflow
javascript tensorflow
asked Nov 20 '18 at 22:01
M. BediM. Bedi
946815
asked Nov 20 '18 at 22:01
M. BediM. Bedi
946815
edited Nov 21 '18 at 19:49
M. Bedi
asked Nov 20 '18 at 22:01
M. BediM. Bedi
946815
asked Nov 20 '18 at 22:01
M. BediM. Bedi
946815
asked Nov 20 '18 at 22:01
M. BediM. Bedi
946815
946815
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
Posenet doesn't track people that way. It just looks for people and creates an array in somewhat random order. In other words, it starts over with every frame, with no knowledge of the previous frames. You would have to write an algorithm of your own.
My suggestion would be to look at the points in the face in each frame and compare them to the points from the previous frame. A naive approach would be to simply find the face in the previous frame that is closest to the face in the current frame. If you needed to do this in a more sophisticated way, you would need do blob tracking where you are actually looking at the colors/composition of each person and making sure they match across frames.
answered Dec 19 '18 at 21:50
Michael ConnorMichael Connor
3,0621819
That is more or less what we ended up doing in JavaScript. It is adequate for now. We are also looking to use ML with sensor data from hardware sensors from Intel & Orbecc. There are plans to make the code and some components open source.
– M. Bedi
Dec 19 '18 at 21:54
add a comment |
Your Answer
StackExchange.ifUsing("editor", function () {
StackExchange.using("externalEditor", function () {
StackExchange.using("snippets", function () {
StackExchange.snippets.init();
});
});
}, "code-snippets");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "1"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53402248%2fissue-with-multi-person-tracking-with-posenet-using-javascript-tensorflow%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
Posenet doesn't track people that way. It just looks for people and creates an array in somewhat random order. In other words, it starts over with every frame, with no knowledge of the previous frames. You would have to write an algorithm of your own.
My suggestion would be to look at the points in the face in each frame and compare them to the points from the previous frame. A naive approach would be to simply find the face in the previous frame that is closest to the face in the current frame. If you needed to do this in a more sophisticated way, you would need do blob tracking where you are actually looking at the colors/composition of each person and making sure they match across frames.
answered Dec 19 '18 at 21:50
Michael ConnorMichael Connor
3,0621819
That is more or less what we ended up doing in JavaScript. It is adequate for now. We are also looking to use ML with sensor data from hardware sensors from Intel & Orbecc. There are plans to make the code and some components open source.
– M. Bedi
Dec 19 '18 at 21:54
add a comment |
Posenet doesn't track people that way. It just looks for people and creates an array in somewhat random order. In other words, it starts over with every frame, with no knowledge of the previous frames. You would have to write an algorithm of your own.
My suggestion would be to look at the points in the face in each frame and compare them to the points from the previous frame. A naive approach would be to simply find the face in the previous frame that is closest to the face in the current frame. If you needed to do this in a more sophisticated way, you would need do blob tracking where you are actually looking at the colors/composition of each person and making sure they match across frames.
answered Dec 19 '18 at 21:50
Michael ConnorMichael Connor
3,0621819
That is more or less what we ended up doing in JavaScript. It is adequate for now. We are also looking to use ML with sensor data from hardware sensors from Intel & Orbecc. There are plans to make the code and some components open source.
– M. Bedi
Dec 19 '18 at 21:54
add a comment |
Posenet doesn't track people that way. It just looks for people and creates an array in somewhat random order. In other words, it starts over with every frame, with no knowledge of the previous frames. You would have to write an algorithm of your own.
My suggestion would be to look at the points in the face in each frame and compare them to the points from the previous frame. A naive approach would be to simply find the face in the previous frame that is closest to the face in the current frame. If you needed to do this in a more sophisticated way, you would need do blob tracking where you are actually looking at the colors/composition of each person and making sure they match across frames.
answered Dec 19 '18 at 21:50
Michael ConnorMichael Connor
3,0621819
Posenet doesn't track people that way. It just looks for people and creates an array in somewhat random order. In other words, it starts over with every frame, with no knowledge of the previous frames. You would have to write an algorithm of your own.
My suggestion would be to look at the points in the face in each frame and compare them to the points from the previous frame. A naive approach would be to simply find the face in the previous frame that is closest to the face in the current frame. If you needed to do this in a more sophisticated way, you would need do blob tracking where you are actually looking at the colors/composition of each person and making sure they match across frames.
answered Dec 19 '18 at 21:50
Michael ConnorMichael Connor
3,0621819
answered Dec 19 '18 at 21:50
Michael ConnorMichael Connor
3,0621819
answered Dec 19 '18 at 21:50
Michael ConnorMichael Connor
3,0621819
answered Dec 19 '18 at 21:50
Michael ConnorMichael Connor
3,0621819
3,0621819
That is more or less what we ended up doing in JavaScript. It is adequate for now. We are also looking to use ML with sensor data from hardware sensors from Intel & Orbecc. There are plans to make the code and some components open source.
– M. Bedi
Dec 19 '18 at 21:54
add a comment |
That is more or less what we ended up doing in JavaScript. It is adequate for now. We are also looking to use ML with sensor data from hardware sensors from Intel & Orbecc. There are plans to make the code and some components open source.
– M. Bedi
Dec 19 '18 at 21:54
That is more or less what we ended up doing in JavaScript. It is adequate for now. We are also looking to use ML with sensor data from hardware sensors from Intel & Orbecc. There are plans to make the code and some components open source.
– M. Bedi
Dec 19 '18 at 21:54
That is more or less what we ended up doing in JavaScript. It is adequate for now. We are also looking to use ML with sensor data from hardware sensors from Intel & Orbecc. There are plans to make the code and some components open source.
– M. Bedi
Dec 19 '18 at 21:54
add a comment |
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53402248%2fissue-with-multi-person-tracking-with-posenet-using-javascript-tensorflow%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
