Face Recognition

説明 Face Recognition

Site Editorial Commentary：Face Recognition can be used as a test framework for face recognition methods

Face Recognition serves as a robust test framework for exploring various face recognition methodologies, integrating Neural Networks with both TensorFlow and Caffe deep learning frameworks. Dive into the world of facial biometrics and unlock the potential of AI-driven face detection on your Android device.

It incorporates a comprehensive suite of preprocessing algorithms to enhance accuracy and performance:

- Grayscale conversion for simplified image analysis
- Precise Cropping to focus on regions of interest
- Eye Alignment to normalize facial orientation
- Gamma Correction for improved contrast
- Difference of Gaussians for edge enhancement
- Canny-Filter for robust edge detection
- Local Binary Pattern for texture-based feature extraction
- Histogram Equalization (requires Grayscale conversion) for enhanced image contrast
- Image Resizing for optimized processing speeds

Choose from a diverse selection of feature extraction and classification methods:

- Eigenfaces coupled with Nearest Neighbour for efficient recognition
- Image Reshaping with Support Vector Machine for powerful classification
- TensorFlow integration with SVM or KNN classifiers
- Caffe integration with SVM or KNN classifiers

Refer to the detailed user manual for comprehensive instructions: https://github.com/Qualeams/Android-Face-Recognition-with-Deep-Learning/blob/master/USER%20MANUAL.md

Currently, the application supports devices with armeabi-v7a architecture and above.

For an optimal recognition experience, rotate your device to the left in recognition mode.

_______________________________________________________________

TensorFlow:

To utilize the TensorFlow Inception5h model, download it from:
https://storage.googleapis.com/download.tensorflow.org/models/inception5h.zip

Copy the "tensorflow_inception_graph.pb" file to "/sdcard/Pictures/facerecognition/data/TensorFlow".

Start with these default settings:
Number of classes: 1001 (irrelevant as the last layer is not used)
Input Size: 224
Image mean: 128
Output size: 1024
Input layer: input
Output layer: avgpool0
Model file: tensorflow_inception_graph.pb
---------------------------------------------------------------------------------------------------------
To utilize the VGG Face Descriptor model, download it from:
https://www.dropbox.com/s/51wi2la5e034wfv/vgg_faces.pb?dl=0

Caution: This model requires at least 3 GB of RAM.

Copy the "vgg_faces.pb" file to "/sdcard/Pictures/facerecognition/data/TensorFlow".

Start with these default settings:
Number of classes: 1000 (irrelevant as the last layer is not used)
Input Size: 224
Image mean: 128
Output size: 4096
Input layer: Placeholder
Output layer: fc7/fc7
Model file: vgg_faces.pb
_______________________________________________________________

Caffe:

To utilize the VGG Face Descriptor model, download it from:
http://www.robots.ox.ac.uk/~vgg/software/vgg_face/src/vgg_face_caffe.tar.gz

Caution: This model requires at least 3 GB of RAM.

Copy the files "VGG_FACE_deploy.prototxt" and "VGG_FACE.caffemodel" to "/sdcard/Pictures/facerecognition/data/caffe".

Start with these default settings:
Mean values: 104, 117, 123
Output layer: fc7
Model file: VGG_FACE_deploy.prototxt
Weights file: VGG_FACE.caffemodel

_______________________________________________________________

The license files can be found here https://github.com/Qualeams/Android-Face-Recognition-with-Deep-Learning/blob/master/LICENSE.txt and here https://github.com/Qualeams/Android-Face-Recognition-with-Deep-Learning/blob/master/NOTICE.txt