Member Avatar for cerezo_mark

Hi i am a College Student and has a thesis of Face Recognition, this thesis was great and i want to have an experience building this thesis. But the problem is i don't know where to start.


I have already studied some of structure approach of this thesis and apparently i discovered that it could be done in Visual C# language, but my problem is i do not how to build it.

Can anyone know how to build it? and somehow what kind of Library or devices etc. i'll be using in order to do this kind of thesis?

I am using right now a Visual Studio 2008.

Any reply would be great from you guys...

Thank you... and GOD Bless...

Edited by cerezo_mark because: n/a
  • 2 Contributors
  • 2 Replies
  • 298 Views
  • 1 Day Discussion Span
  • Latest Post Latest Post by cerezo_mark

Dani AI

Generated

— good project. Break the problem into two clear parts: face detection (find faces in frames) and face recognition (match a found face to a person). ’s webcam pointer is a fine way to start getting live frames; begin by proving you can reliably capture and save labelled images from a camera before moving on to recognition.

For C# libraries consider an OpenCV wrapper (Emgu CV) for mature detection/recognition building blocks, and Accord.NET or AForge.NET for simple capture and machine-learning utilities. Emgu CV exposes Haar/LBP cascades for detection and common recognizers (PCA/Eigenfaces, LDA/Fisherfaces, LBPH). If you must stay on Visual Studio 2008 check each library’s .NET target (some require newer frameworks), otherwise upgrade to a newer VS for easier dependency management.

A practical workflow to include in your thesis:

  • Data capture: record 20–50 labelled images per subject with varied lighting/pose; keep consistent background where possible. Obtain consent and note privacy handling.
  • Preprocessing: detect face, align/crop, convert to grayscale, histogram-equalize, resize (e.g., 100x100).
  • Algorithms: implement baseline PCA (eigenfaces), LDA (fisherfaces), and LBPH. Use simple classifier (nearest neighbour) then try SVM or k-NN for comparison.
  • Evaluation: use 70/30 split or k-fold cross-validation; report accuracy, confusion matrix, and ROC/AUC. Document hyperparameters and failure cases.

Quick Emgu CV sketch (pseudocode):

using Emgu.CV;
using Emgu.CV.Structure;

var capture = new Capture(0);
var frame = capture.QueryFrame().ToImage<Bgr,byte>();
var gray = frame.Convert<Gray,byte>();

var cascade = new CascadeClassifier("haarcascade_frontalface_default.xml");
var faces = cascade.DetectMultiScale(gray, 1.1, 3);

foreach (var r in faces) {
  var face = gray.Copy(r).Resize(100,100, Emgu.CV.CvEnum.Inter.Cubic);
  recognizer.Train(face, label);
}

Troubleshooting tips: if no faces are detected, check cascade file path, lighting, image size, and detection parameters (scale and minNeighbors). For a thesis, compare methods, be transparent about dataset limits, and include ethical/privacy discussion.

Recommended Answers

All 2 Replies

Member Avatar for pseudorandom21

Hi i am a College Student and has a thesis of Face Recognition, this thesis was great and i want to have an experience building this thesis. But the problem is i don't know where to start.


I have already studied some of structure approach of this thesis and apparently i discovered that it could be done in Visual C# language, but my problem is i do not how to build it.

Can anyone know how to build it? and somehow what kind of Library or devices etc. i'll be using in order to do this kind of thesis?

I am using right now a Visual Studio 2008.

Any reply would be great from you guys...

Thank you... and GOD Bless...

Is this what you want?

If not, you'll have to be more specific.

Member Avatar for cerezo_mark

Thank you sir for giving some LINK, and i apologize if i did not tell exactly what i want to do with it.

But specifically sir i want to build a Face Recognition Software just like this http://www.youtube.com/watch?v=6nRLFXkPrKs.


Thank you again sir for replying.. i appreciate it very much..

Edited by cerezo_mark because: n/a
Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.