GauTure's Paper Accepted for Publication in the Internationally Recognized Academic Journal Scientific Data

On October 10, 2023, GauTure's R&D team's paper "An annotated grain kernel image database for visual quality inspection" was accepted for publication in Scientific Data, a journal under Nature, which is a well-known international academic journal.

  Acceptance Notification Email

  Nature Publishing enjoys a reputation for excellence and is widely recognized in the academic and research community.Scientific Data journals have a recent impact factor of 8.5 in 2023, are listed in Region 1 of the international JCR (Journal Citation Reports), and are included in the SCI (Science Citation Index).

  Overview of the core of the thesis

  In the paper, we present a multi-region, multi-grain, machine vision-based database called GrainSet, designed for visual quality checking of grain particles. The database contains over 350,000 single-grain images with information on image, quality, and imperfect grains. The grain particles used in the study consisted of four cereal grains, wheat, maize, sorghum, and rice, which were obtained from more than 20 major grain-producing regions in five countries around the world. We use a self-customized multi-grain collection device that is equipped with multiple high-resolution optical sensor units to efficiently and automatically collect multi-grains. In addition, we test the classification performance of multiple deep learning models on GrainSet. We believe that GrainSet will contribute to future research in areas such as assisting in grain quality inspection, providing guidance on grain storage and trade, and smart agricultural applications.

  Part of the paper

  Journal Introduction

  Scientific Data is Nature's online journal dedicated to publishing scientifically valuable datasets and research that promotes the sharing and reuse of scientific data. It covers all areas of the natural sciences, medicine, engineering and social sciences. With the goal of improving the discoverability, accessibility, comprehensibility, and reusability of data, thereby advancing science. According to Scopus' CiteScore metrics, Scientific Data is ranked #1 out of 152 journals in the broad category of statistics and data in the social sciences.