English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Due to the growth of video data on Internet, automatic video analysis hasgained a lot of attention from academia as well as companies such as Facebook,Twitter and Google. In this paper, we examine the robustness of video analysisalgorithms in adversarial settings. Specifically, we propose targeted attackson two fundamental classes of video analysis algorithms, namely videoclassification and shot detection. We show that an adversary can subtlymanipulate a video in such a way that a human observer would perceive thecontent of the original video, but the video analysis algorithm will return theadversary's desired outputs.  We then apply the attacks on the recently released Google Cloud VideoIntelligence API. The API takes a video file and returns the video labels(objects within the video), shot changes (scene changes within the video) andshot labels (description of video events over time). Through experiments, weshow that the API generates video and shot labels by processing only the firstframe of every second of the video. Hence, an adversary can deceive the API tooutput only her desired video and shot labels by periodically inserting animage into the video at the rate of one frame per second. We also show that thepattern of shot changes returned by the API can be mostly recovered by analgorithm that compares the histograms of consecutive frames. Based on ourequivalent model, we develop a method for slightly modifying the video frames,in order to deceive the API into generating our desired pattern of shotchanges. We perform extensive experiments with different videos and show thatour attacks are consistently successful across videos with differentcharacteristics. At the end, we propose introducing randomness to videoanalysis algorithms as a countermeasure to our attacks.

Attacking Automatic Video Analysis Algorithms: A Case Study of Google Cloud Video Intelligence API