facial_recognition.md

Masked Face Recognition

📝 How does a facial recognition system works?

1.Face detection: first, the system has to identify the part of the image or the video that represents the face. 2. Pre-processing: the data has to be transformed into a normalized, monolith format (the images have to be of the same resolution, levels of zoom, brightness, orientation, etc.). It is also often referred to as feature normalization. 3. Feature extraction: the system has to extract meaningful data from the facial images, identifying the most relevant bits of data and ignoring all of the “noise.” It’s also referred to as encoding. 4. Face recognition: the actual process of matching unique data features to each individual. The underlying principle here is called object classification.

📝Deep face recognition using imperfect facial data

1. Paper Link
2. Studies the performance of Deep Methods for partial face recognition
3. Experiments are based on CNN based architecture along with the pre-trained VGG-Face model through which we extract features for machine learning.
4. Cosine similarity and SVM has been used as classifiers
5. Several experiments conducted using both adding individual face parts and not adding them to the training set
6. The experiment claims to reach 90% accuracy in both the cases
7. Adding partial faces images to the training dataset boosted the accuracy
8. Cosine Similarity seems to be better measure than SVMs for the respective task

---

📝Occlusion Robust Face Recognition Based on Mask Learning With Pairwise Differential Siamese Network

1. Link to the article
2. Human visual system pays attention to the non-occluded facial areas for recognition (and ignores the occluded areas), the paper propose to discard feature elements that have been corrupted by occlusions
3. To learn the correspondence between occluded facial regions and corrupted feature elements, we develop a novel pairwise differential siamese network (PDSN) structure with a mask generator module that takes pairwise images (one is a clean face and the other is an occluded one of the same identity) as input.
4. Our mask generator will identify those feature elements that are harmful for the recognition as well as far from its genuine values as corrupted ones.
5. Given the feature of a face image with block bj occluded, denoted as f(xj ), how to learn a mask generator Mθ whose output is multiplied with the f(xj ) to mask out those corrupted elements.
6. The core mask generator module Mθ in our PDSN is expected to output a mask whose element is a real value in [0, 1] and is multiplied with the input contaminated feature to diminish its corrupted elements: ˜f(xij) = Mθ(·)f(xij)
7. A fixed mask is extracted from every trained mask generator Mθ and build a dictionary accordingly.
8. After getting the mask, the final features can be calculated. The similarity score between two samples can be calculated using Cosine Similarity.

📝 Efficient Detection of Occlusion prior to Robust Face Recognition Link

1. The proposed approach consists of first detecting and segmenting occluded parts (e.g., sunglasses/scarves) and then applying face recognition on the non-occluded facial regions
2. the presence of occlusion is first analysed in the patch-level using Gabor wavelets, PCA and SVM.
3. Then we segment the occluded part more precisely from the other facial regions by a generalized Potts model Markov random field (GPM-MRF)
4. After the computation of an occlusion mask indicating which pixel in a face image is occluded, we propose a variant of local Gabor binary pattern histogram sequences (LGBPHS) [11] to efficiently represent occluded faces by excluding features extracted from the occluded pixels

📝 Occlusion invariant Face Recognition Link