phillity's profile - activity

(2) becomes a video-based approach if the features of multiple sequential frames are used together in order to make a si

Based on your description, something like (2) is what you're going to need. If you deal with static images you are going

Based on your description, something like (2) is what you're going to need. If you deal with static images you are going

If you google "face anti-spoofing" or "liveness detection," you will see many articles and research papers which try to

If you google "face anti-spoofing" or "liveness detection," you will see many articles and research papers which try to

If you google "face anti-spoofing" or "liveness detection," you will see many articles and research papers which try to

If you google "face anti-spoofing" or "liveness detection," you will see many articles and research papers which try to

If you google "face anti-spoofing" or "liveness detection," you will see many articles and research papers which try to

If you google "face anti-spoofing" or "liveness detection," you will see many articles and research papers which try to

If you google "face anti-spoofing" or "liveness detection," you will see many articles and research papers which try to

If you google "face anti-spoofing" or "liveness detection," you will see many articles and research papers which try to

If you google "face anti-spoofing" you will see many articles and research papers which try to accomplish this task. The

If you google "face anti-spoofing" you will see many articles and research papers which try to accomplish this task. The

CVPR 2019 papers on this topic: http://openaccess.thecvf.com/content_CVPR_2019/papers/Liu_Deep_Tree_Learning_for_Zero-Sh

In addition to the ML course by Andrew Ng, I would also recommend his DL courses on Coursera. These DL courses (particul

No, confidence should be a Euclidean distance. If you get a negative value that would be an indication there is somethin

No, confidence should be a Euclidean distance. If you get a negative value that would be an indication there is somethin

Ya I think the logic is correct. Since the facial bounding boxes are different, facial landmarks are found using differe

@LBerger Ya, unless the facial landmark points seem wrong or are really different, I would think it's okay The facial b

@LBerger Ya, unless the facial landmark points seem wrong or are really different, I would think it's okay The facial b

LBPHFaceRecognizer performs facial identification using KNN. LBPH is used to extract facial features. During predictio

@oja if you found the answer helpful, please upvote and accept it. Good luck!

In a facial authentication project I used opencv.js for accessing the user's camera and allowing users to take and submi

In VS Code, press Ctrl+Shift+p, then make sure you have the correct python environment selected in which you installed O

When creating a facial landmark dataset, a researcher has to go through many facial images and use tools to annotate the

When creating a facial landmark dataset, a researcher has to go through many facial images and use tools to annotate the

Please see this webpage -- "The network training started with randomly initialized weights and used a structured metric

Please see this webpage -- "The network training started with randomly initialized weights and used a structured metric

I used this tensorflow code for head pose estimation a while back and it worked well for me

Here is the aforementioned pyimagesearch link. It uses both OpenCV and Dlib

I am new to opencv and am trying to convert this line of matlab code to java using opencv 3.1:

• user_feature_matrix_normalize = abs(round(user_feature_matrix.*100)); //matlab code

I think I figured out multiplication and this is what I have so far:

• Mat user_feature_matrix_normalize = new Mat(); //java opencv code
• Scalar alpha =new Scalar(100);
• Core.multiply(user_feature_matrix, alpha, user_feature_matrix_normalize);

Are there any opencv functions I can use to round and take the absolute value of each element of the mat? Looking for any suggestions on how I should try to accomplish this

Hi everyone,

I am going to try to set up a fingerprint authentication system. I am looking for an openly available fingerprint database to use for feature extraction testing and authentication/rejection testing.

I have found a couple databases available online but they all only have one fingerprint for each subject (right thumb, right index, left index, etc.). This has posed me problems when trying to test authentication as there is no second fingerprint that a test fingerprint can be authenticated as. Can anyone point me to a database where each subject has multiple fingerprints for each finger?

Hi everyone,

I have been working through the steps of fingerprint matching presented in an answer here: http://answers.opencv.org/question/63...

Some of the links provided in the answer are now dead and I cannot find examples online. I have figured out how to perform some steps, but there are a few I am struggling with.

For the blockwise gabor filtering, can someone provide me an example or the parameters I should use? (All the images I am using are size 388x374)

EDIT:

I have begun working on this again and made some progress. I have obtained the orientation map. I think there should be some way to improve my results though, as they don't seem as good as the result in the referenced answer.

Currently, I first perform CLAHE to increase image contrast. Then, I am obtaining the gradient magnitude and direction for each pixel. Next, I split the image into non-overlapping windows. I average the directions within each window to find the local direction for the windows and plot a corresponding lines. Is there some improvement I can make to this process to better my results? Is this the correct way to get "block-wise gradient direction" as the referenced answer states?

My next step will be to enhance the image using block-wise gabor filters. Should they be applied to each window using the window's average direction I found? Can anyone further explain the answer's step: "Enhance the normalize image by block-wise gabor filter according to block-wise orientation map then binarize the fingerprint"

Any tips will be very appreciated!

Initialize OpenCL Runtime Hi everyone I am making a web app which uses some OpenCV functions on the backend. The first

Here is the 2014 CVPR paper it is based on "Face alignment at 3000 fps via regressing local binary features"

First the training images are organized into a single training dataset matrix where each image is a single column/row (d

LDA is just a generalization of FDA. See wiki link and stack exchange link. So yes, the final result should be the same

You could try KMeans along with Connect Components