news.ucr.edu Open in urlscan Pro
2600:9000:2251:aa00:15:6601:6480:93a1  Public Scan

Form analysis
1 forms found in the DOM

Name: gsc-search-form — /results

<form id="audience-search-form" name="gsc-search-form" action="/results">
  <div class="gsc-modal-body">
    <div class="grid-container full">
      <div class="grid-x grid-padding-x grid-padding-y align-center-middle text-center">
        <div class="gsc-text-wrapper">
          <span class="gsc-welcome-text">Let us help you with your search</span>
        </div>
        <div class="cell">
          <label id="searchbox-label" class="hidden" aria-label="Enter your Search Criteria.">Enter your Search Criteria.</label>
          <input type="text" maxlength="255" id="audience-search" name="q" value="" aria-labelledby="searchbox-label" placeholder="Search for...">
        </div>
        <div class="cell medium-4">
          <button class="button gsc-modal-button-submit" type="submit" name="search-by" value="all">Search All UCR</button>
        </div>
        <div class="cell medium-4">
          <button class="button gsc-modal-button-submit" type="submit" name="search-by" value="news.ucr.edu">Search This Site</button>
        </div>
        <div class="cell medium-4">
          <button type="button" class="button gsc-modal-button-close" data-close="">Cancel</button>
        </div>
      </div>
    </div>
  </div>
</form>

Text Content

COVID-19 and return to campus information. 

×
Skip to main content

University of California, Riverside

Search

UC Riverside


NEWS



Search
 * Home
 * Latest Articles
   * Arts / Culture
   * Athletics
   * Business
   * Health
   * Science / Technology
   * Social Science / Education
   * Students
   * University
 * Calendar
 * Experts
 * In the News
 * Info for Media
 * Inside UCR
 * UCR Magazine
   * Winter 2022
   * Fall 2021

Breadcrumb
 *  
 * News
 * May 3rd, 2022
 * New method detects deepfake videos with up to 99% accuracy

Follow US:


May 3, 2022


NEW METHOD DETECTS DEEPFAKE VIDEOS WITH UP TO 99% ACCURACY

Two-pronged technique detects manipulated facial expressions and identity swaps


Author: Holly Ober
May 3, 2022
Share This:
FacebookTwitterLinkedInEmailPrintFriendly

Computer scientists at UC Riverside can detect manipulated facial expressions in
deepfake videos with higher accuracy than current state-of-the-art methods. The
method also works as well as current methods in cases where the facial identity,
but not the expression, has been swapped, leading to a generalized approach to
detect any kind of facial manipulation. The achievement brings researchers a
step closer to developing automated tools for detecting manipulated videos that
contain propaganda or misinformation.

Developments in video editing software have made it easy to exchange the face of
one person for another and alter the expressions on original faces. As
unscrupulous leaders and individuals deploy manipulated videos to sway political
or social opinions, the ability to identify these videos is considered by many
essential to protecting free democracies. Methods exist that can detect with
reasonable accuracy when faces have been swapped. But identifying faces where
only the expressions have been changed is more difficult and to date, no
reliable technique exists.

“What makes the deepfake research area more challenging is the competition
between the creation and detection and prevention of deepfakes which will become
increasingly fierce in the future. With more advances in generative models,
deepfakes will be easier to synthesize and harder to distinguish from real,”
said paper co-author Amit Roy-Chowdhury, a Bourns College of Engineering
professor of electrical and computer engineering.




First and second columns show the original im- ages and manipulated ones
respectively. The black and white images in the third column are corresponding
bi- nary GT masks. Predicted masks (column 4) and generated CAMs (column 5) for
manipulated images from Face2Face (row 1,2,3) and Neural-Textures (row 4,5,6)
dataset. (Mazaheri & Roy-Chowdhury, 2022)

The UC Riverside method divides the task into two components within a deep
neural network. The first branch discerns facial expressions and feeds
information about the regions that contain the expression, such as the mouth,
eyes, or forehead, into a second branch, known as an encoder-decoder. The
encoder-decoder architecture is responsible for manipulation detection and
localization.

The framework, called Expression Manipulation Detection, or EMD, can both detect
and localize the specific regions within an image that have been altered.

“Multi-task learning can leverage prominent features learned by facial
expression recognition systems to benefit the training of conventional
manipulation detection systems. Such an approach achieves impressive performance
in facial expression manipulation detection,” said doctoral student Ghazal
Mazaheri, who led the research.

The benchmark datasets for facial manipulation are based on expression and
identity swap.  One transfers the expressions of a source video onto a target
video without changing the identity of the person in the target video. The other
swaps two identities in a single video.  

Experiments on two challenging facial manipulation datasets show EMD has better
performance in detection of not only facial expression manipulations but also
identity swaps. EMD accurately detected 99% of the manipulated videos.

The paper, “Detection and Localization of Facial Expression Manipulations,” was
presented at the 2022 Winter Conference on Applications of Computer Vision and
is available here.

Share this Article
FacebookTwitterLinkedInEmailPrintFriendly


MEDIA CONTACTS

HOLLY OBER

Senior Public Information Officer
Email (951) 827-5893
Tags
Bourns College Of Engineering
Department Of Computer Science And Engineering
Amit Roy-Chowdhury
deepfakes
algorithms
Science / Technology


RELATED ARTICLES


Science / Technology


UCR ECOLOGISTS WORK TOWARD POST-FIRE REBIRTH OF HEALTHY LANDSCAPES   

Science / Technology


INSIGHTS FROM ALGAE GENES UNLOCK MYSTERIES OF PLANT GROWTH AND HEALTH

Science / Technology


HOW DRONES CAN HELP DAIRY FARMS MANAGE METHANE EMISSIONS

Science / Technology


CANDY-COATED PILLS COULD PREVENT PHARMACEUTICAL FRAUD

Search
University of California, Riverside

900 University Ave.
Riverside, CA 92521

Tel: (951) 827-1012

 * UCR Library
 * Campus Status
 * Campus Store
 * Career Opportunities
 * Diversity
 * Maps and Directions
 * Visit UCR


UNIVERSITY OF CALIFORNIA, RIVERSIDE

900 University Ave.
Riverside, CA 92521

tel: (951) 827-0000
email: webmaster@ucr.edu

Find Us




RELATED LINKS

 * UCR News Archive
 * Science Today
 * Servicio de Información en Español
 * UC Agricultural and Natural Resources news
 * UC Newsroom
 * Available Feeds
 * Creator State Podcast

Follow Us:

           
 * Privacy and Accessibility
 * Terms and Conditions
 * © 2022 Regents of the University of California

✓
Thanks for sharing!
AddToAny
More…

Search ×
Let us help you with your search
Enter your Search Criteria.
Search All UCR
Search This Site
Cancel