@inproceedings{oai:kyutech.repo.nii.ac.jp:00005677,
 author = {Nishi,  Toshiki and Kurogi, Shuichi and 黒木, 秀一 and Matsuo, Kazuya and 松尾, 一矢},
 book = {2017 IEEE Symposium Series on Computational Intelligence (SSCI)},
 month = {Dec},
 note = {This paper presents a method for grading fruits and vegetables by means of using RGB-D (RGB and depth) images and convolutional neural network (CNN). Here, we focus on grading according to the size of objects. First, the method transforms positions of pixels in RGB image so that the center of the object in 3D space is placed at the position equidistant from the focal point by means of using the corresponding depth image. Then, with the transformed RGB images involving equidistant objects, the method uses CNN for learning to classify the objects or fruits and vegetables in the images for grading according to the size, where the CNN is structured for achieving both size sensitivity for grading and shift invariance for reducing position error involved in images. By means of numerical experiments, we show the effectiveness and the analysis of the present method., The 2017 IEEE Symposium Series on Computational Intelligence (IEEE SSCI 2017), November 27 to December 1, 2017, Honolulu, Hawaii, USA},
 publisher = {IEEE},
 title = {Grading Fruits and Vegetables Using RGB-D Images and Convolutional Neural Network},
 year = {2017},
 yomi = {クロギ, シュウイチ and マツオ, カズヤ}
}