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/

Global: Zendy Tools annual

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

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

An intuition on human segmentation is that when a human is moving in a video,the video-context (e.g., appearance and motion clues) may potentially inferreasonable mask information for the whole human body. Inspired by this, basedon popular deep convolutional neural networks (CNN), we explore a very-weaklysupervised learning framework for human segmentation task, where only animperfect human detector is available along with massive weakly-labeled YouTubevideos. In our solution, the video-context guided human mask inference and CNNbased segmentation network learning iterate to mutually enhance each otheruntil no further improvement gains. In the first step, each video is decomposedinto supervoxels by the unsupervised video segmentation. The superpixels withinthe supervoxels are then classified as human or non-human by graph optimizationwith unary energies from the imperfect human detection results and thepredicted confidence maps by the CNN trained in the previous iteration. In thesecond step, the video-context derived human masks are used as direct labels totrain CNN. Extensive experiments on the challenging PASCAL VOC 2012 semanticsegmentation benchmark demonstrate that the proposed framework has alreadyachieved superior results than all previous weakly-supervised methods withobject class or bounding box annotations. In addition, by augmenting with theannotated masks from PASCAL VOC 2012, our method reaches a new state-of-the-artperformance on the human segmentation task.

Learning to Segment Human by Watching YouTube