遇见数据集

Image Dataset for: Effect of Camera Choice on Image-Classification Inference

收藏
Zenodo2025-01-11 更新2026-05-26 收录
官方服务:

资源简介:

This is the full image data set for the following research paper: Brown, J., Nguyen, A., & Raj, N. (2025). Effect of Camera Choice on Image-Classification Inference. Applied Sciences, 15(1), 246. https://doi.org/10.3390/app15010246 The dataset comprises images of cups and mugs captured using various cameras, backgrounds, camera-to-target distances and image resolutions. There is a folder for each of 6 camera models. Within each folder, the image files are named using the following convention: <camera_model>_<cup_or_mug_identifier>_<orientation>_<camera-to-target_distance>_<image_resolution>.jpg The <camera_model> field has possible values: Angetube962A, EMEETSmartCamC960, LogitechC270, LogitechC920e, MicrosoftLifeCam, RazerKiyoX The <cup_or_mug_identifier> field has possible values Cup1, Cup2, Mug1, Mug2, Mug3 The <orientation> field has possible values Orientation1, Orientation2, Orientation3 to reflect the different image backgrounds. The <camera-to-target_distance> field has possible values 20cm, 30cm, 40cm, 50cm, 60cm The <image_resolution> field has various values depending on camera support, please see the above paper for details. In the above paper, different terms are used: 'vessel' to indicate a cup or mug, and 'scenario' to indicate an orientation/background. The mapping between the terms used in the paper and the fields used in the file names are as follows: Term used in paper <cup_or_mug_identifier> field in file name Vessel 1 Mug1 Vessel 2 Cup1 Vessel 3 Mug2 Vessel 4 Cup2 Vessel 5 Mug3 Term used in paper <orientation> field in file name Scenario 1 Orientation3 Scenario 2 Orientation2 Scenario 3 Orientation1

提供机构:
Zenodo
创建时间:
2025-01-11
二维码
社区交流群
二维码
科研交流群
商业服务