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 Plus annual

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

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Weakly supervised object detection (WSOD) is a challenging task, in whichimage-level labels (e.g., categories of the instances in the whole image) areused to train an object detector. Many existing methods follow the standardmultiple instance learning (MIL) paradigm and have achieved promisingperformance. However, the lack of deterministic information leads to partdomination and missing instances. To address these issues, this paper focuseson identifying and fully exploiting the deterministic information in WSOD. Wediscover that negative instances (i.e. absolutely wrong instances), ignored inmost of the previous studies, normally contain valuable deterministicinformation. Based on this observation, we here propose a negativedeterministic information (NDI) based method for improving WSOD, namelyNDI-WSOD. Specifically, our method consists of two stages: NDI collecting andexploiting. In the collecting stage, we design several processes to identifyand distill the NDI from negative instances online. In the exploiting stage, weutilize the extracted NDI to construct a novel negative contrastive learningmechanism and a negative guided instance selection strategy for dealing withthe issues of part domination and missing instances, respectively. Experimentalresults on several public benchmarks including VOC 2007, VOC 2012 and MS COCOshow that our method achieves satisfactory performance.

Absolute Wrong Makes Better: Boosting Weakly Supervised Object Detection via Negative Deterministic Information