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

Deep learning has made great strides for object detection in images. Thedetection accuracy and computational cost of object detection depend on thespatial resolution of an image, which may be constrained by both the camera andstorage considerations. Compression is often achieved by reducing eitherspatial or amplitude resolution or, at times, both, both of which havewell-known effects on performance. Detection accuracy also depends on thedistance of the object of interest from the camera. Our work examines theimpact of spatial and amplitude resolution, as well as object distance, onobject detection accuracy and computational cost. We develop aresolution-adaptive variant of YOLOv5 (RA-YOLO), which varies the number ofscales in the feature pyramid and detection head based on the spatialresolution of the input image. To train and evaluate this new method, wecreated a dataset of images with diverse spatial and amplitude resolutions bycombining images from the TJU and Eurocity datasets and generating differentresolutions by applying spatial resizing and compression. We first show thatRA-YOLO achieves a good trade-off between detection accuracy and inference timeover a large range of spatial resolutions. We then evaluate the impact ofspatial and amplitude resolutions on object detection accuracy using theproposed RA-YOLO model. We demonstrate that the optimal spatial resolution thatleads to the highest detection accuracy depends on the 'tolerated' image size.We further assess the impact of the distance of an object to the camera on thedetection accuracy and show that higher spatial resolution enables a greaterdetection range. These results provide important guidelines for choosing theimage spatial resolution and compression settings predicated on availablebandwidth, storage, desired inference time, and/or desired detection range, inpractical applications.

Understanding the Impact of Image Quality and Distance of Objects to Object Detection Performance