Attention! Beginners, Theory: Challenging terms and Methods for Facial Features Recognition
Last Updated on July 20, 2023 by Editorial Team
Author(s): Surya Govind
Originally published on Towards AI.
Theory Explained ( Computer Vision ): Pattern recognition ( Facial Recognition ) with Challenges and possible methods
Energy-saving tip folk: Better start with OpenCV-C++/ Python, https://docs.opencv.org/2.4/modules/contrib/doc/facerec/facerec_tutorial.html is a library of Python bindings designed to solve computer vision problems. Later you can keep progressing with neural networks.
In 1960 Woody gave the first semi-automated facial recognition technique. This technique required some features such as mouth, nose, eyes, and nose on the image. After that, face recognition techniques came in the picture. In 2001 onwards this biometric technique increased rapidly according to the below figure. The main goal of this article is to aware beginners about constantly increasing scientific interest in face recognition biometric technique. The below analysis has been performed by applying the keyword “face recognition”.
A. Aging :
Aging is an inevitable natural process during the lifetime of a person as compared to other facial variations. Aging effect can be observed under the main three unique characteristics:
1) The aging is uncontrollable: It cannot be advanced or even delayed and it is slow and irreversible.
2) Personalized aging Signs: Every human passes through different aging patterns. And these rely on his or her genes and many other factors, such as health, food, region, and weather conditions.
3) The aging signs depend on time: The face of a person at age will affect all older faces, but unaffected at a younger age.
B. Partial Occlusion :
You see, the above image is pretty fair with partial occlusion, isn’t it! , but hey wait, what about this one?
It goes like, Occlusion refers to natural or artificial obstacles in an image. It can be a local region of the face along with different objects such as sunglasses, scarves, hands, and hair. They are generally called partial occlusions. Partial occlusions correspond to any occluding object. And the occlusion less than 50% of the face is considered to be a partial occlusion. The approaches to face recognition with partial occlusion are classified into the following three categories:
(1) Part Based Methods.
(2) Feature-based methods and
(3) Fractal-Based Methods.
Many areas of image processing have been impacted by partial occlusion such as recognition by ear is occluded due to earrings. Occlusion affects the performance of a system when people deceive it either by the use of sunglasses, scarves, veil or by placing mobile phones or hands in front of faces. In some cases, other factors like shadows due to extreme illumination also act as occluding factors. Further, local approaches are used to deal with the problem of partially occluded faces which divide the faces into different parts. However, this problem can be overcome by eliminating some of the features which create trouble while accurate recognition in the image. Mostly local methods are based on feature analysis, in which the best possible features are detected and then they are combined. Another approach that can be applied for this purpose is near a holistic approach in which occlude features, traits and characters are eradicated and the rest of the face is used as valuable information. Different techniques are being developed by the researchers to cope up with this problem. Oh gosh! pretty much to understand, right?
C. Pose Invariance :
Pose variance is yet another hurdle in achieving a successful face recognition system. People pose differently every time they take a picture. There is no standardized rule for taking a pose. Therefore, it makes more difficult to distinguish and recognize the faces from images with varying poses. Pose variations degrade the performance of the facial features. Also, many systems work under inflexible imaging conditions and as a result, it affects the quality of gallery images. The methods dealing with variation in the pose can be divided into two kinds i.e. multi-view face recognition and face recognition across pose. Multi-view face recognition can be considered as an annexure of frontal face recognition in which the gallery image of every pose is considered. On the other hand, across a pose in face recognition, yield face with a pose that has never been exposed before to a recognition system. There are some other methods and approaches that are being used to tackle the similar problem of face recognition. Furthermore, variance and changes in the pose can be divided into three classes, namely:
(1) General algorithms.
(2) Two-dimensional methods for face recognition and
(3) Three-dimensional models.
D. Illuminations :
Illumination is an observable property and effect of light. It may also refer to the lightning effect or the use of light sources. Global illuminations are algorithms that have been used in 3D computer graphics. Illumination variation also badly affects the face recognition system. Thus it has been turned an area of attention for many researchers. However, it becomes a tedious task to recognize one or more persons from still or video images. But it can be quite easy to extract desired information from images when they are taken under a controlled environment along with the uniform background. Also, three methods that can be implemented to deal with the illumination problem. They are gradient, grey level, and face reflection field estimation techniques. Grey level transformation technique carries out in-depth mapping with a non- linear or linear function. Gradient extraction approaches are used to extract edges of an image at a grey level. Illumination is a factor that heavily affects the performance of the recognition system obtained via face images or videos. These techniques are developed to suppress the effect of illumination.
METHODS USED IN FACE RECOGNITION :
Face recognition is the sub-area of pattern recognition research and technology. Firstly we take input images from the standard cameras and this image is known as the input image. Secondly, pre-processing is performed to improve the image quality and to reduce the noise in the taken images. The next phase is face segmentation where the face part is a cropped from the human body and background of the image. In the next phase, feature extraction is done after segmenting the input image to a very good level and relevant features are identified which can help in distinguishing that person to other people. In the last phase, a template is generated for enrollment and matching purposes. Finally, matching is done at the biometric system for identification of the authenticated person, if there is a match then the output will come in the form of user accepted otherwise user rejected. Face recognition systems can be divided into different categories:
A. Knowledge-based method :
The knowledge-based techniques are based on the geometry of the face and arrangement of the facial features. These knowledge-based methods describe the shape, size, and texture. Some methods describe a few other characteristics of facial features such as head, eyebrows, eyes, nose, and chin. The major problem of these techniques is that they do not perform well because of different types of pose or head orientations as shown in my implementation results of V-J algorithm in the figure:
B. Feature invariant approaches:
The main aim of feature invariant techniques is to find structural features of the human face even with lighting conditions varying. Different types of structural features such as facial local features, skin color, shape, and texture are used. These methods are very sensitive to illumination, occlusion, the existence of skin color regions, and adjacent faces as shown in the figure:
C. Template-based methods:
Template-based methods are sensitive to pose, scale and shape variation of the human body. Deformable template methods have been proposed to deal with such variations of pose, scale, and shape of the body. Template-based methods using the elastic models include shape parameters as well as intensity information of facial features.
D. Appearance-based methods:
The appearance-based methods used for face detection are Eigenfaces, Linear Discriminant Analysis, Neural Networks, Support Vector Machine, and Hidden Markov Models. With the help of these methods, the whole image is scanned and image regions are identified as face or no-face. During template matching, appearance-based methods will come in the picture.
So hey!, Finally, face recognition has remained a striving area for researchers for many years. In this article, a comprehensive study was performed over different face recognition methods and challenging terms. After detailed analysis, it revealed that PCA is the best-suited technique when the dimension of features is higher for original face images. The matter of fact is, Human behavior has emerged recently as a promising research area for scholars that should be exploited in the future. Finally, it is concluded that still there remains a gap in terms of study in the face recognition system that requires to be filled to improve its accuracy and efficiency.
Congratulations folk !, You have made the first step in face recognition. I would happily hear your questions and providing the best suitable answers.
HAPPY LEARNING FAM!
Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.
Published via Towards AI