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

Despite being widely used as a performance measure for visual detectiontasks, Average Precision (AP) is limited in (i) reflecting localisationquality, (ii) interpretability and (iii) robustness to the design choicesregarding its computation, and its applicability to outputs without confidencescores. Panoptic Quality (PQ), a measure proposed for evaluating panopticsegmentation (Kirillov et al., 2019), does not suffer from these limitationsbut is limited to panoptic segmentation. In this paper, we propose LocalisationRecall Precision (LRP) Error as the average matching error of a visual detectorcomputed based on both its localisation and classification qualities for agiven confidence score threshold. LRP Error, initially proposed only for objectdetection by Oksuz et al. (2018), does not suffer from the aforementionedlimitations and is applicable to all visual detection tasks. We also introduceOptimal LRP (oLRP) Error as the minimum LRP Error obtained over confidencescores to evaluate visual detectors and obtain optimal thresholds fordeployment. We provide a detailed comparative analysis of LRP Error with AP andPQ, and use nearly 100 state-of-the-art visual detectors from seven visualdetection tasks (i.e. object detection, keypoint detection, instancesegmentation, panoptic segmentation, visual relationship detection, zero-shotdetection and generalised zero-shot detection) using ten datasets toempirically show that LRP Error provides richer and more discriminativeinformation than its counterparts. Code available at:https://github.com/kemaloksuz/LRP-Error

One Metric to Measure them All: Localisation Recall Precision (LRP) for Evaluating Visual Detection Tasks