Tag Archive: Video fragmentation

Presentation of InVID verification technologies at JUDNM-T

InVID verification technologies at JUDNM-T

The InVID project and the InVID verification technologies, were presented on a workshop focusing on tools for content verification. The workshop was co-organized by CERTH/InVID and the Media Informatics Lab of the Journalism & Mass Media Dept. of the Aristotle University of Thessaloniki, in collaboration with the Open Knowledge Greece Foundation and the Journalists’ Union of Daily Newspapers of Macedonia – Thrace (JUDNM-T). It took place on Wednesday 6th of June 2018, in the conference centre of JUDNM-T, in Thessaloniki, Greece. The audience of the workshop (approximately 40 people), was comprised by journalists, academics and Master students of the Journalism & Mass Media Dept.

After the introductory part of the event, which was made by Prof. Andreas Veglis, the workshop started by an overall presentation of the InVID project by Dr. Vasileios Mezaris, the project co-ordinator. Dr. Mezaris explained the project’s motivation with the help of some indicative examples of fake news. Following, he described the project’s goals and overall concept for meeting these goals, and briefly discussed the developed integrated technologies for newsworthy media collection and verification. The given presentation can be seen below.

Subsequently, Mr. Evlampios Apostolidis from CERTH-ITI, gave a talk on fake news that rely on video reuse. In particular, Mr. Apostolidis presented some examples of such fakes, discussed the most common approaches of journalists to deal with this type of fake news and explained how the InVID technologies can assist journalists to identify the original video source. For this, he presented the developed web application for video fragmentation and reverse keyframe search, that enables the users/journalists to perform reverse video search at the fragment-level, and detect previously published versions of the same video online. The given presentation can be found below.

Following, Dr. Symeon Papadopoulos from CERTH-ITI, discussed the current status on the creation and spread of fake news and rumors via social networks and listed a number of credibility signals that could allow a user (i.e. a journalist or a media verification expert) to evaluate the veracity of a news item. Finally, we presented the functionalities and characterists of the developed context analysis and aggregation toolkit that has been developed in InVID and is integrated in the InVID Verification Plugin, and highlighted the creation (and future public release) of a constantly extendable Fake Video Corpus, that can support the current and future reseach on this field.

The first part of the workshop ended with a presentation focusing on image forensics analysis, made by Dr. Markos Zampoglou from CERTH-ITI. Dr. Zampoglou explained the existing challenges related to the detection of tampering of digital images, highlighting the fact that the sharing of these images via social media platforms entails a re-compression process that usually erases most of the digital traces that could assist the identification of manipulations. Based on this observation, he stressed the need to analyse the most earliest version of a digital image, and discussed the funtionalities of the REVEAL Media Verificaiton Assistant that is integrated in the InVID Verification Plugin. The given presentation can be seen in the following.

The second part of the workshop was a hands-on session, where the participants of the workshop had the opportunity to get activelly involved in a step-by-step process containing:

  • the installation of the InVID Verification Plugin for both Chrome and Firefox users (made by Evlampios Apostolidis);
  • a brief explanation of all different components integrated in the plugin (given by Evlampios Apostolidis);
  • a real-time debunking of a couple of examples of fake news based on video re-use (performed by Evlampios Apostolidis);
  • a real-time debunking of a couple of examples of fake news based on context change (performed by Olga Papadopoulou, also from CERTH-ITI);
  • a real-time debunking of a couple of examples of fake news based on image tampering (performed by Markos Zampoglou).

The aforementioned examples were used for demonstration purposes, while an additional set of examples was provided to the participants of the workshop, aiming to motivate them to discover the functionalities, capabilities and limitations of the different components of the plugin, and send us their feedback (via the integrated survey) concerning their experience with this tool.

French minister attends InVID demo

French minister attend InVID demo

The newly appointed French minister of Higher Education, Research and Innovation, Mrs Frédérique Vidal, visited on Thursday June 8th the InVID booth at Futur en Seine Digital Festival 2017 (see central image above) where she was shown prototypes aiming to debunk fake videos. In particular, Mrs Vidal was shown how the fragmentation service, extracting relevant keyframes from a video, could help to debunk quickly a viral video shared during the French presidential elections by extremists xenophobic Facebook pages, claiming that a migrant seeking free universal healthcare assaulted a nurse and other employees in a public hospital.

The image reverse search demonstrated in a few clicks that the video was in fact taken one month earlier in Novgorod, Russia, and was showing a drunk man hitting there the hospital employees. Once the video was debunked within the CrossCheck initiative led by First Draft News and Google News Lab on the French election (check the report about the debunking here), it was removed from Facebook but after being seen more than seven millions times.

The image reverse search demonstrated in a few clicks that the video was in fact taken one month earlier in Novgorod, Russia, and was showing a drunk man hitting there the hospital employees. Once the video was debunked within the CrossCheck initiative led by First Draft News and Google News Lab on the French election (check the report about the debunking here), it was removed from Facebook but after being seen more than seven millions times.

With the help of the video fragmentation and reverse keyframe search tool was shown that the video was in fact taken one month earlier in Novgorod, Russia, presenting a drunk man hitting the hospital employees.

The recent presidential election raised deep concern in France about the fake news spreading on social networks. French president Emmanuel Macron himself complained shortly before his victory that «the social networks hurt me a lot» through attacks from hardliners extremists.

InVID project at Futur en Seine digital festival in Paris

InVID project at Futur en Seine digital festival in Paris

The early prototypes of InVID will be exposed at the Futur en Seine digital festival in Paris, at the Grande halle de la Villette (practical info available here), from the 8th to the 10th of June, where professionals and the public will be able to see and test how journalists can debunk fake videos on social networks with examples taken from recent breaking and social media emerging news.

Partner AFP will demo the InVID Discovery platform (a.k.a. the InVID Multimodal Analytics Dashboard), the InVID Verification Application and an InVID browser plugin: a verification toolbox soon to be released in open source.

The browser plugin, tested over the last few weeks by the video and social media team at AFP, allows to quickly debunk a fake video by extracting thumbnails from the corresponding web platform, or by fragmenting the video into keyframes (see screenshot bellow) before searching those images on a reverse image search engine like Google Images to retrieve previous copies of the same video if any available. And this works for Facebook, Youtube, Twitter or any video file the journalist chooses to upload to InVID platform.

InVID-keyframes

Recently when the news of a Manila resort Casino in Philippines being attacked on 1st of June 2017 evening broke, a fake video (first screenshot bellow)  started to circulate on Twitter, claiming to be a raw footage of the attack from a CCTV camera while as debunk by an AFP social media journalist, it was a copy of previous videos on another attack perpetrated at a hotel in Suriname at the end of December 2011 (second screenshot bellow).

Fake video about a robbery take place in a casino in Manila

Fake video claiming to show a robbery at a Manila resort Casino on 1th of June 2017.

The real video

The original video showing an attack at the Savanah hotel in Suriname on 27th-28th of December 2011.