Face Verification using MLKIT Vision

Question

Hi Iam developing an android app using MLKIT Vision for face detection. - I can able to detect the user faces using mlkit vision and store the detected faces in internal storage. Now i want to verify the user faces with stored faces like. If a user opens the app starting capturing the user face and i want to verify with the stored face. How to verify the two faces are identical or not using the MLKIT vision?

Rajanikant Hawaldar · Answer

No Use OpenCV for this comparison task

Ravikant Sahu · Answer

you can do it like this // Step 1: Store contour val storedcontours: List = // List of face contour stored in internal storage // Step 2: Capture User Face val capturedImage: Bitmap = // Captured user's face image val capturedFacecontour: FloatArray = // Compute contour for the captured face using ML Kit // Step 3: Compare contour for (storedcontour in storedcontours) { val similarityScore = calculateSimilarityScore(capturedFacecontour storedcontour) if (similarityScore > threshold) { // Face match found user is verified break } }

Saravanan Ganesan · Answer

To verify if two faces are identical using ML Kit Vision: Capture the user's face using ML Kit Vision for face detection. Compare the captured face with the stored face using face recognition techniques or similarity metrics like cosine similarity. If the similarity score exceeds a certain threshold consider the faces as identical. Please note that ML Kit Vision primarily focuses on face detection not face recognition. For face recognition you might need to use a dedicated face recognition library or implement a deep learning model.