<head> <meta http-equiv="Content-Type" content="text/html"> <!--<title>Table</title> --> <style type="text/css"> BODY { background: white; /* */ } TABLE { /* width: 300px; /* */ border: 1; border-collapse: collapse; /* */ border: 1px solid ; /* white*/ } TD { height: 70; /*width: 70; */ min-width: 100; padding: 3px; /* */ border: 1px solid black; /* */ text-align: left; /* */ background: DarkSalmon; z-index: 1; } TH { /* min-width: 50px; */ /* min-height: 100px; */ position: -webkit-sticky; position: sticky; top: 0; padding: 3px; /* */ border: 1px solid black; /* */ text-align: left; /* */ background: rgb(151, 151, 151); } p.vertical{ -webkit-transform: rotate(-90deg); -moz-transform: rotate(-90deg); -ms-transform: rotate(-90deg); -o-transform: rotate(-90deg); transform: rotate(-90deg); /*height:260px; */ } </style> </head> <body> <div> <table> <caption> Luna SDK mobile models and dependencies</caption> <thead> <tr> <th height="200"><center><p class="horizontal"><b>Feature name and <br> description</b></p></center></th> <th height="200"><center><p class="horizontal"><b>Required algorithms and used structures</b></p></center></th> <th height="200"><center><p class="horizontal"><b>Best Shot <br>Selection</b></p></center></th> <th height="200"><center><p class="horizontal"><b>Face Features<br> Extraction</b></p></center></th> <th height="200"><center><p class="horizontal"><b>Face <br>Recognition</b></p></center></th> </tr> </thead> <!-- --> <tfoot> <tr> <td colspan="30" style="background-color: White"> <p align="left" > The <device> placeholder shows that the model can be used with different devices (e. g., CPU, GPU). <br> Name of the model includes corresponding postfix (e. g., *_cpu.plan, *_gpu.plan). <br> See "Hardware requirements" section in "FaceEngine Handbook" for information about devices supported by NN.</p> <p align="left" > Example of table using - if user wants to find best shots in his application, he needs to look at the "Best Shot Selection"<br> column and see what network models are required in order to run such functionality.<br>The required models and structures are <span style="color: DarkSeaGreen"> marked with green</span>. <br>Algorigthms and structures <span style="color: DarkSalmon"> marked with red</span> are not used for the functionality.</p> </td> </tr> </tfoot> <!-- --> <tbody> <tr> <td style="background-color: Silver">Face detection NN version 2<br>First, second and third NN <br>should be launched consequently</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IDetector<br><br>Plan(s):<br> FaceDet_2_first_<device>.plan<br> FaceDet_2_second_<device>.plan<br> FaceDet_2_third_<device>.plan</th> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> </tr> <tr> <td style="background-color: Silver">BestShotQuality estimation<br>Performs head pose and AGS estimations</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IBestShotQualityEstimator<br><br>Plan(s):<br>ags_angle_estimation_flwr_<device>.plan</td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> </tr> <tr> <td style="background-color: Silver">Image quality estimation <br>(light, dark, blur, etc.)</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IQualityEstimator<br><br>Plan(s):<br> model_subjective_quality_<version>_<device>.plan</td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> </tr> <tr> <td style="background-color: Silver">Image glasses estimation <br>(light, dark, blur, etc.)</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IGlassesEstimator<br><br>Plan(s):<br> glasses_estimation_flwr_<version>_<device>.plan</td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> </tr> <tr> <td style="background-color: Silver">Head pose estimation (deprecated)</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IHeadPoseEstimator<br><br>Plan(s):<br>angle_estimation_flwr_<device>.plan</td> <td style="background-color: rgb(188, 168, 143)"></td> <td style="background-color: rgb(188, 168, 143)"></td> <td style="background-color: rgb(188, 168, 143)"></td> </tr> <tr> <td style="background-color: Silver">AGS estimation (deprecated)</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IAGSEstimator<br><br>Plan(s):<br>ags_estimation_flwr_<device>.plan </td> <td style="background-color: rgb(188, 168, 143)"></td> <td style="background-color: rgb(188, 168, 143)"></td> <td style="background-color: rgb(188, 168, 143)"></td> </tr> <tr> <td style="background-color: Silver">LivenessOneShotRGB estimation<br>Predicts whether the person's face is real or fake (photo, printed image)</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>ILivenessOneShotRGBEstimator<br><br>Plan(s):<br>oslm_v3_model_<version>_<device>.plan</td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> </tr> <tr> <td style="background-color: Silver">Extraction (NN version 54 mobilenet)<br>(Available in Complete distribution package only!) </td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IDescriptor, IDescriptorBatch, IDescriptorExtractor<br><br>Plan(s):<br>cnn54m_<device>.plan</th> <td></td> <td></td> <td style="background-color: DarkSeaGreen"></td> </tr> <tr> <td style="background-color: Silver">Eyes state estimation (eyes status and <br>keypoints for irises and eyelids)</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IEyeEstimator<br><br>Plan(s):<br>eyes_estimation_flwr8_<device>.plan</td> <td></td> <td style="background-color: DarkSeaGreen"></td> <td></td> </tr> <tr> <td style="background-color: Silver">Eyes state <br>(lightened version, returns eyes state only)</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IEyeEstimator<br><br>Plan(s):<br>eye_status_estimation_flwr_<device>. <td></td> <td style="background-color: DarkSeaGreen"></td> <td></td> </tr> <div> <table> <caption> Luna SDK functionality</caption> <thead> <tr> <th height="100"><center><p class="horizontal"><b>Functionality</b></p></center></th> <th height="100"><center><p class="horizontal"><b>Feature name and <br> description</b></p></center></th> <th height="100"><center><p class="horizontal"><b>Required algorithms and used structures</b></p></center></th> <th height="100"><center><p class="horizontal"><b>Face detection NN<br> versions 1 and 2</b></p></center></th> <th height="100"><center><p class="horizontal"><b>Face detection NN <br> version 3</b></p></center></th> </tr> </thead> <tr> <td style="background-color: Silver">MedicalMaskDetection</td> <td style="background-color: Silver">Medical mask estimation</td> <td style="background-color: rgba(240, 238, 238, 0.952)">Structure(s):<br>IMedicalMaskEstimator<br><br>Plan(s):<br>mask_clf_<version>_<device>.plan</td> <td style="background-color: DarkSeaGreen"></td> <td style="background-color: DarkSeaGreen"></td> </tr> </table> </div> <br> <div style="background-color: rgba(240, 238, 238, 0.952); border: 1px solid #ccc; padding-left: 10px;"> <p align="left" > The table above includes separately licensed feature. </p> <p align="left" > Additional features required for the functionality utilization are shown in the columns to the right. </p> </div> <!-- !!!!!! --> <tr class="even"> <td style="background-color: white; height: 5" colspan="15" ></td> </tr> </tbody> </table> </div> </html>