Images of Pigs at Different Stages From a Group of Shandong Farms

(1)A set of images of pigs in different stages from Shandong, China. After,years of image acquisition, more than one images, videos and video frame extraction images of piglets (including piglets around 50 days old and 70-80 days old), fattening pigs, sows and male breeding pigs were collected.(2) The dataset contains 63 images for piglets around 50 days, 139 images for those between 70 and 80 days, 91 images for fattening pigs at 160 days and 26 images for sows.(3) The dataset contains 1 video for piglets at about 50 days, 3 videos for those at 70-80 days, 5 videos for fattening pigs at 160 days, 4 videos for sows, and 4 videos for male breeding pigs.(4) Pig dataset from pig farms in Shandong for visual feature-based classification.(5) The dataset of multi-class problem is called multi-class problem because it has visual features such as reducing the chance of overfitting, improving accuracy, and reducing the training time of the model.

（1）本数据集涵盖采自中国山东省不同生长阶段生猪的影像数据。历经多年影像采集工作，团队收集了仔猪（包含约50日龄与70~80日龄仔猪）、育肥猪、母猪及种公猪的多幅图像、视频及视频抽帧图像。（2）数据集包含约50日龄仔猪图像63幅、70~80日龄仔猪图像139幅、160日龄育肥猪图像91幅以及母猪图像26幅。（3）数据集包含约50日龄仔猪视频1段、70~80日龄仔猪视频3段、160日龄育肥猪视频5段、母猪视频4段以及种公猪视频4段。（4）本数据集源自山东各生猪养殖场，用于基于视觉特征的生猪分类任务。（5）该多分类数据集之所以被称为多分类问题数据集，是因为其相关视觉特性可降低模型过拟合概率、提升分类准确率并缩短模型训练时长。