How do I crop faces with a neural network?












2












$begingroup$


I am looking to build a facial recognition system and realized I could probably pump up accuracy by first cropping the faces. I know I can use a Haar Cascade classifier to do this but would prefer to use a deep learning approach, as that is my current area of focus. To do this do I simply create a CNN with inputs being an image and outputs being the X, Y, Width, Height? Or is there more to it? I have tried this approach with an Adam optimizer and around 600 pictures but results haven't been great. Any ideas on how to improve this?










share|improve this question







New contributor




user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.







$endgroup$












  • $begingroup$
    You may want to use pixel-wise classification, fully convolutional networks. YOLO based approaches can be considered too.
    $endgroup$
    – Media
    25 mins ago
















2












$begingroup$


I am looking to build a facial recognition system and realized I could probably pump up accuracy by first cropping the faces. I know I can use a Haar Cascade classifier to do this but would prefer to use a deep learning approach, as that is my current area of focus. To do this do I simply create a CNN with inputs being an image and outputs being the X, Y, Width, Height? Or is there more to it? I have tried this approach with an Adam optimizer and around 600 pictures but results haven't been great. Any ideas on how to improve this?










share|improve this question







New contributor




user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.







$endgroup$












  • $begingroup$
    You may want to use pixel-wise classification, fully convolutional networks. YOLO based approaches can be considered too.
    $endgroup$
    – Media
    25 mins ago














2












2








2





$begingroup$


I am looking to build a facial recognition system and realized I could probably pump up accuracy by first cropping the faces. I know I can use a Haar Cascade classifier to do this but would prefer to use a deep learning approach, as that is my current area of focus. To do this do I simply create a CNN with inputs being an image and outputs being the X, Y, Width, Height? Or is there more to it? I have tried this approach with an Adam optimizer and around 600 pictures but results haven't been great. Any ideas on how to improve this?










share|improve this question







New contributor




user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.







$endgroup$




I am looking to build a facial recognition system and realized I could probably pump up accuracy by first cropping the faces. I know I can use a Haar Cascade classifier to do this but would prefer to use a deep learning approach, as that is my current area of focus. To do this do I simply create a CNN with inputs being an image and outputs being the X, Y, Width, Height? Or is there more to it? I have tried this approach with an Adam optimizer and around 600 pictures but results haven't been great. Any ideas on how to improve this?







python deep-learning keras






share|improve this question







New contributor




user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.











share|improve this question







New contributor




user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.









share|improve this question




share|improve this question






New contributor




user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.









asked 35 mins ago









user3647894user3647894

111




111




New contributor




user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.





New contributor





user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.






user3647894 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.












  • $begingroup$
    You may want to use pixel-wise classification, fully convolutional networks. YOLO based approaches can be considered too.
    $endgroup$
    – Media
    25 mins ago


















  • $begingroup$
    You may want to use pixel-wise classification, fully convolutional networks. YOLO based approaches can be considered too.
    $endgroup$
    – Media
    25 mins ago
















$begingroup$
You may want to use pixel-wise classification, fully convolutional networks. YOLO based approaches can be considered too.
$endgroup$
– Media
25 mins ago




$begingroup$
You may want to use pixel-wise classification, fully convolutional networks. YOLO based approaches can be considered too.
$endgroup$
– Media
25 mins ago










1 Answer
1






active

oldest

votes


















0












$begingroup$

You are looking for a fully convolution neural network trained for image segmentation. The architecture is typically convolutions followed by de-convolutional layers with out pooling.





The forward inference serves to extract information with convolutions. The de-convolutions serve to resize the tensors to the size of the inputs height and width. A fully connected layer then classifies each pixel from the input. Deconvolutions demonstration.



Unfortunately you need pixel wise labels to train a network to do this. You could look for a pretrained model that already segments faces from images or you could grind out hand labeling for a 1000 images and use transfer learning (there are online tools for doing this such as Label box.





share









$endgroup$













    Your Answer





    StackExchange.ifUsing("editor", function () {
    return StackExchange.using("mathjaxEditing", function () {
    StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
    StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
    });
    });
    }, "mathjax-editing");

    StackExchange.ready(function() {
    var channelOptions = {
    tags: "".split(" "),
    id: "557"
    };
    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: false,
    noModals: true,
    showLowRepImageUploadWarning: true,
    reputationToPostImages: null,
    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
    });


    }
    });






    user3647894 is a new contributor. Be nice, and check out our Code of Conduct.










    draft saved

    draft discarded


















    StackExchange.ready(
    function () {
    StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f44456%2fhow-do-i-crop-faces-with-a-neural-network%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









    0












    $begingroup$

    You are looking for a fully convolution neural network trained for image segmentation. The architecture is typically convolutions followed by de-convolutional layers with out pooling.





    The forward inference serves to extract information with convolutions. The de-convolutions serve to resize the tensors to the size of the inputs height and width. A fully connected layer then classifies each pixel from the input. Deconvolutions demonstration.



    Unfortunately you need pixel wise labels to train a network to do this. You could look for a pretrained model that already segments faces from images or you could grind out hand labeling for a 1000 images and use transfer learning (there are online tools for doing this such as Label box.





    share









    $endgroup$


















      0












      $begingroup$

      You are looking for a fully convolution neural network trained for image segmentation. The architecture is typically convolutions followed by de-convolutional layers with out pooling.





      The forward inference serves to extract information with convolutions. The de-convolutions serve to resize the tensors to the size of the inputs height and width. A fully connected layer then classifies each pixel from the input. Deconvolutions demonstration.



      Unfortunately you need pixel wise labels to train a network to do this. You could look for a pretrained model that already segments faces from images or you could grind out hand labeling for a 1000 images and use transfer learning (there are online tools for doing this such as Label box.





      share









      $endgroup$
















        0












        0








        0





        $begingroup$

        You are looking for a fully convolution neural network trained for image segmentation. The architecture is typically convolutions followed by de-convolutional layers with out pooling.





        The forward inference serves to extract information with convolutions. The de-convolutions serve to resize the tensors to the size of the inputs height and width. A fully connected layer then classifies each pixel from the input. Deconvolutions demonstration.



        Unfortunately you need pixel wise labels to train a network to do this. You could look for a pretrained model that already segments faces from images or you could grind out hand labeling for a 1000 images and use transfer learning (there are online tools for doing this such as Label box.





        share









        $endgroup$



        You are looking for a fully convolution neural network trained for image segmentation. The architecture is typically convolutions followed by de-convolutional layers with out pooling.





        The forward inference serves to extract information with convolutions. The de-convolutions serve to resize the tensors to the size of the inputs height and width. A fully connected layer then classifies each pixel from the input. Deconvolutions demonstration.



        Unfortunately you need pixel wise labels to train a network to do this. You could look for a pretrained model that already segments faces from images or you could grind out hand labeling for a 1000 images and use transfer learning (there are online tools for doing this such as Label box.






        share











        share


        share










        answered 4 mins ago









        J_HeadsJ_Heads

        1473




        1473






















            user3647894 is a new contributor. Be nice, and check out our Code of Conduct.










            draft saved

            draft discarded


















            user3647894 is a new contributor. Be nice, and check out our Code of Conduct.













            user3647894 is a new contributor. Be nice, and check out our Code of Conduct.












            user3647894 is a new contributor. Be nice, and check out our Code of Conduct.
















            Thanks for contributing an answer to Data Science Stack Exchange!


            • 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.


            Use MathJax to format equations. MathJax reference.


            To learn more, see our tips on writing great answers.




            draft saved


            draft discarded














            StackExchange.ready(
            function () {
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f44456%2fhow-do-i-crop-faces-with-a-neural-network%23new-answer', 'question_page');
            }
            );

            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







            Popular posts from this blog

            How to label and detect the document text images

            Vallis Paradisi

            Tabula Rosettana