search

nwesem / mtcnn_facenet_cpp_tensorRT

Face Recognition on NVIDIA Jetson (Nano) using TensorRT
GNU General Public License v3.0
200 stars 72 forks source link

facenet is not accurate #16

Closed xudh1991 closed 3 years ago

xudh1991 commented 3 years ago

I have run through the whole process, but the facenet is not accurate. I tested the accuracy of face recognition, but the accuracy rate is only 20%. I suspect there are some errors between the image and the model, such as normalization, color change, I tried to find the error but no success I looked at the feature vector, the whole value is very small, most are less than 0.1 I need help, please give me some suggestions， thanks

gugumituo commented 3 years ago

Hi,have you solve the problem?I also meet the same issue. I found when i put two same pictures to the model,the distance is 0; But i put a cropped image from the old one,the distance becomes a big value,which is bigger than 1000. I don't know why,please help me!

lnxguyz commented 3 years ago

I am facing the same issue, pls help

xudh1991 commented 3 years ago

Hi,have you solve the problem?I also meet the same issue. I found when i put two same pictures to the model,the distance is 0; But i put a cropped image from the old one,the distance becomes a big value,which is bigger than 1000. I don't know why,please help me!

xudh1991 commented 3 years ago

I've given up the project. I suggest giving up