Every Smile is Unique: Landmark-Guided Diverse Smile Generation

نویسندگان

  • Wei Wang
  • Xavier Alameda-Pineda
  • Dan Xu
  • Elisa Ricci
  • Nicu Sebe
چکیده

Each smile is unique: one person surely smiles in different ways (e.g. closing/opening the eyes or mouth). Given one input image of a neutral face, can we generate multiple smile videos with distinctive characteristics? To tackle this one-to-many video generation problem, we propose a novel deep learning architecture named Conditional MultiMode Network (CMM-Net). To better encode the dynamics of facial expressions, CMM-Net explicitly exploits facial landmarks for generating smile sequences. Specifically, a variational auto-encoder is used to learn a facial landmark embedding. This single embedding is then exploited by a conditional recurrent network which generates a landmark embedding sequence conditioned on a specific expression (e.g. spontaneous smile). Next, the generated landmark embeddings are fed into a multi-mode recurrent landmark generator, producing a set of landmark sequences still associated to the given smile class but clearly distinct from each other. Finally, these landmark sequences are translated into face videos. Our experimental results demonstrate the effectiveness of our CMM-Net in generating realistic videos of multiple smile expressions.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Role of smile correction in mineral detection on hyperion data

This work aims to extract the mineralogical constituents of the Lahroud Hyperion scene situated in the NW of Iran. Like the other push-broom sensors, Hyperion images suffer from spectral distortions, namely the smile effect. The corresponding spectral curvature is defined as an across-track wavelength shift from the nominal central wavelength, and alters the pixel spectra. The common “column me...

متن کامل

Spontaneous Smile Detection with Application of Landmark Points Supported by Visual Indications

When automatic recognition of emotion became feasible, novel challenges has evolved. One of them is the recognition whether a presented emotion is genuine or not. In this work, a fully automated system for differentiation between spontaneous and posed smiles is presented. This solution exploits information derived from landmark points, which track the movement of fiducial elements of face. Addi...

متن کامل

Smile Analyzer: A Software Package for Analyzing the Characteristics of the Speech and Smile

Taking into account the factors related to lip-tooth relationships in orthodontic diagnosis and treatment planning is of prime importance. Manual quantitative analysis of facial parameters on photographs during smile and speech is a difficult and time-consuming job. Since there is no comprehensive and user-friendly software package, we developed a software program called "Smile Analyzer" in the...

متن کامل

Evaluation of Relationship between Smile Line and Age

Smile line is one of the most important element of the smile esthetics. In an attractive smile, the maxillary anterior incisal curve is parallel to the lower lip. With the changes in face, lips and teeth muscles, which are caused by aging, this element (smile line) also changes. Aging causes smile line to move from its parallel position and in complete smile, formulate different shapes relative...

متن کامل

Landmark-free smile intensity estimation

Facial expression analysis is an important field of research, mostly because of the rich information faces can provide. The majority of works published in the literature have focused on facial expression recognition and so far estimating facial expression intensities have not gathered same attention. The analysis of these intensities could improve face processing applications on distinct areas,...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:
  • CoRR

دوره abs/1802.01873  شماره 

صفحات  -

تاریخ انتشار 2018