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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

This paper proposes an algorithm for recognizing slab identification numbersin factory scenes. In the development of a deep-learning based system, manuallabeling to make ground truth data (GTD) is an important but expensive task.Furthermore, the quality of GTD is closely related to the performance of asupervised learning algorithm. To reduce manual work in the labeling process,we generated weakly annotated GTD by marking only character centroids. Whereasbounding-boxes for characters require at least a drag-and-drop operation or twoclicks to annotate a character location, the weakly annotated GTD requires asingle click to record a character location. The main contribution of thispaper is on selective distillation to improve the quality of the weaklyannotated GTD. Because manual GTD are usually generated by many people, it maycontain personal bias or human error. To address this problem, the informationin manual GTD is integrated and refined by selective distillation. In theprocess of selective distillation, a fully convolutional network is trainedusing the weakly annotated GTD, and its prediction maps are selectively used torevise locations and boundaries of semantic regions of characters in theinitial GTD. The modified GTD are used in the main training stage, and apost-processing is conducted to retrieve text information. Experiments werethoroughly conducted on actual industry data collected at a steelmaking factoryto demonstrate the effectiveness of the proposed method.

Selective Distillation of Weakly Annotated GTD for Vision-based Slab Identification System