Hi,
I'm implementing facial recognition using Harcascade classifier. At first, I'm creating the dataset with the input images and then training the images using the classifier . Then, I'm performing recognition in live video using the same classifier and LBPH Face Recognizer. The recognition detects faces with bounding boxes, but the results are inaccurate.
Facing the following issues:
- Face is recognized but returns wrong "id" for the recognized face.
- Unknown faces are also recognized with "id's", but there is no such face in dataset and no id is assigned to it.
- Face is recognized only at a shorter distance from camera.
At times in longer distances, face is detected but not recognised
I also tried the same implementation with lbpcascade classifier and ended up with the same above issues. Researched on this issue but could not find any positive solution for this. Is there any other method that could produce good accurate results? Kindly help me out to fix these issues.
Thanks in advance, Preeti