The Elements of End-to-end Deep Face Recognition: A Survey of Recent Advances

Face recognition is one of the most popular and long-standing topics in computer vision. With the recent development of deep learning techniques and large-scale datasets, deep face recognition has made remarkable progress and been widely used in many real-world applications. Given a natural image or video frame as input, an end-to-end deep face recognition system outputs the face feature for recognition. To achieve this, a typical end-to-end system is built with three key elements: face detection, face alignment, and face representation. The face detection locates faces in the image or frame. Then, the face alignment is proceeded to calibrate the faces to the canonical view and crop them with a normalized pixel size. Finally, in the stage of face representation, the discriminative features are extracted from the aligned face for recognition. Nowadays, all of the three elements are fulfilled by the technique of deep convolutional neural network. In this survey article, we present a comprehensive review about the recent advance of each element of the end-to-end deep face recognition, since the thriving deep learning techniques have greatly improved the capability of them. To start with, we present an overview of the end-to-end deep face recognition. Then, we review the advance of each element, respectively, covering many aspects such as the to-date algorithm designs, evaluation metrics, datasets, performance comparison, existing challenges, and promising directions for future research. Also, we provide a detailed discussion about the effect of each element on its subsequent elements and the holistic system. Through this survey, we wish to bring contributions in two aspects: first, readers can conveniently identify the methods which are quite strong-baseline style in the subcategory for further exploration; second, one can also employ suitable methods for establishing a state-of-the-art end-to-end face recognition system from scratch.

A Python toolbox for Automatic Face Alignment (AFA)

We describe AFA, an open-source Python package for automating the most common step in the preparation of facial stimuli for behavioral and neuro-imaging experiments – spatial alignment of faces (https://github.com/SourCherries/auto-face-align ). Face alignment is also important in the analysis of image statistics, and as a preprocessing step for machine learning. Automation of face alignment via AFA provides a reliable and efficient alternative to the very common practice of manual image-editing in graphics editors like Photoshop. As an open-source Python package, AFA encourages a clear and transparent specification of experimental method. AFA is based on facial landmark detection that is powered by the reliable and open-source DLIB library; and critical alignment code based on Generalized Procrustes Analysis (GPA) has been extensively unit-tested. AFA documentation and modularity provides opportunity for the modification and extensibility of AFA by the scientific community. As examples, we include functions for automatically generating image apertures that conceal areas outside the inner face; for image morphing between facial identities; and for shape-based averaging of facial identity. All of these are examples of stimulus preparation that have previously required manual landmark selection.

3D Face Alignment Method Applied to Face Recognition

Presentation of 3D Face Alignment method with application to Face Recognition

3D Face Alignment Method Applied to Face Recognition

Presentation of 3D Face Alignment method with application to Face Recognition

Novel Completion Designs in Horizontal Well with Transverse Hydraulic Fractures for Enhanced Recovery in Unconventional or Tight Reservoirs

Low recovery, 2 to 15%, in unconventional plays (including tight reservoirs and source rocks) has long been recognized as a business deterrent. The industry applies enhanced oil recovery (EOR) techniques, along with hydraulic fractures in tight/unconventional plays, to improve the recovery. To maximize matrix sweep, the fractures are aligned in a face-to-face assembly. Such an arrangement can be achieved using vertical or longitudinal hydraulic fracture on horizontal wells, but these, generally, do not provide as effective reservoir contact (hydraulic fracture surface area) as horizontal wells with multistage transverse hydraulic fractures. The multistage transverse hydraulic fracture, however, comes at the costs of conformance issues with early water breakthrough from short-circuiting and inability to achieve fracture face-to-fracture face alignment of the injection and production fractures. The vast majority of wells drilled in unconventional plays are in the transverse configuration; hence, there is a need for an optimal solution for transverse fractures combined with improved oil recovery (IOR)/EOR approaches. In this work, we introduce the multistage enhanced recovery (MS-ER) techniques that enable face-to-face alignment for optimal enhanced hydrocarbon recovery/IOR/EOR in horizontal wells with multistage transverse fractures, thereby enabling optimal recovery and mitigating the key risk of fracture short-circuiting.