Open bhachauk opened 4 years ago
Some problem. Does Anybody have progress? Accuracy is only about 16% for one-shot case. Thanks
I have analyzed this one shot learning technique with various transfer learning models... I feel like it seems learning with small set of data not worthy even doing one shot learning (It really doing bad with real time test set).... May be with more data (with large learning time and resource) It will be possible to get decent results...
I did small analysis for face_recognition task with a best i ever tested face feature extractor keras-vggface even compared with dlib. Reference analysis.
Here i made a post about this Reality Of One Shot Learning For Face Recognition
Hi, Thanks for your work. I have used the model
nn4.small2.lrn.h5
for extracting the features after re-runningKeras-openface-convertion.ipynb
.I got worst accuracy while applied with minimum / euclidean distance classifier for one shot face authentication (one image per class) compared with dlib feature extractor.
Am I doing right ? Does the model work for one shot learning or not ?
Thanks and regards. Bhanuchander