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

Are existing object detection methods adequate for detecting text and visualelements in scientific plots which are arguably different than the objectsfound in natural images? To answer this question, we train and compare theaccuracy of various SOTA object detection networks on the PlotQA dataset. Atthe standard IOU setting of 0.5, most networks perform well with mAP scoresgreater than 80% in detecting the relatively simple objects in plots. However,the performance drops drastically when evaluated at a stricter IOU of 0.9 withthe best model giving a mAP of 35.70%. Note that such a stricter evaluation isessential when dealing with scientific plots where even minor localisationerrors can lead to large errors in downstream numerical inferences. Given thispoor performance, we propose minor modifications to existing models bycombining ideas from different object detection networks. While thissignificantly improves the performance, there are still 2 main issues: (i)performance on text objects which are essential for reasoning is very poor, and(ii) inference time is unacceptably large considering the simplicity of plots.To solve this open problem, we make a series of contributions: (a) an efficientregion proposal method based on Laplacian edge detectors, (b) a featurerepresentation of region proposals that includes neighbouring information, (c)a linking component to join multiple region proposals for detecting longertextual objects, and (d) a custom loss function that combines a smooth L1-losswith an IOU-based loss. Combining these ideas, our final model is very accurateat extreme IOU values achieving a mAP of 93.44%@0.9 IOU. Simultaneously, ourmodel is very efficient with an inference time 16x lesser than the currentmodels, including one-stage detectors. With these contributions, we enablefurther exploration on the automated reasoning of plots.

A Systematic Evaluation of Object Detection Networks for Scientific Plots