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 methods have recently exhibited impressive performance inobject detection. However, such methods needed much training data to achievehigh recognition accuracy, which was time-consuming and required considerablemanual work like labeling images. In this paper, we automatically preparetraining data using robots. Considering the low efficiency and high energyconsumption in robot motion, we proposed combining robotic in-hand observationand data synthesis to enlarge the limited data set collected by the robot. Wefirst used a robot with a depth sensor to collect images of objects held in therobot's hands and segment the object pictures. Then, we used a copy-pastemethod to synthesize the segmented objects with rack backgrounds. The collectedand synthetic images are combined to train a deep detection neural network. Weconducted experiments to compare YOLOv5x detectors trained with imagescollected using the proposed method and several other methods. The resultsshowed that combined observation and synthetic images led to comparableperformance to manual data preparation. They provided a good guide onoptimizing data configurations and parameter settings for training detectors.The proposed method required only a single process and was a low-cost way toproduce the combined data. Interested readers may find the data sets andtrained models from the following GitHub repository: github.com/wrslab/tubedet

Automatically Prepare Training Data for YOLO Using Robotic In-Hand Observation and Synthesis