Ultrasound of Dairy Cows' Udder (Mammary Gland) & Sonogram Echotexture Analysis

This public dataset contains:

1. A folder with B-mode sonograms (3072 in total) of the bovine mammary gland.
2. A .csv file with the daily milk yield on measurement days of the dairy cows participating in this study.
3. A .csv file with the feature set deriving from echotexture analysis of the B-mode sonograms.
4. A "Data_Interpretation" file, explaining how to interpret and use the data.

Research Hypothesis:
Ultrasonography is being utilized in the examination of dairy animals’ mammary gland for more than three decades. Recently, echotexture analysis has been employed by some researchers to gain deeper information from mammary gland's sonograms and investigate pathological conditions. In our research article "A Deep Learning Algorithm Predicts Milk Yield and Production Stage of Dairy Cows utilizing Ultrasound Echotexture Analysis of the Mammary Gland" we hypothesized that a deep neural network could utilize features from echotexture analysis of mammary gland sonograms to obtain predictions about the cows’ daily milk yield and production stage.

Notable Findings:
A deep neural network can utilize features from mammary sonograms’ echotexture analysis to predict a cows’ daily milk yield and production stage.
Our research findings are presented in detail in our research article "A Deep Learning Algorithm Predicts Milk Yield and Production Stage of Dairy Cows utilizing Ultrasound Echotexture Analysis of the Mammary Gland".

Data Interpretation:
Text file “Data_Interpretation” includes the information needed to interpret and use the data. An extensive description about the data collection methodology is provided in our research article "A Deep Learning Algorithm Predicts Milk Yield and Production Stage of Dairy Cows utilizing Ultrasound Echotexture Analysis of the Mammary Gland".

本公开数据集包含以下内容：
1. 一个包含牛乳腺B型超声图像（B-mode sonograms，共计3072张）的文件夹。
2. 一份记录参与本研究的奶牛各测量日日产奶量的CSV文件。
3. 一份包含基于B型超声图像回声纹理分析（echotexture analysis）提取的特征集的CSV文件。
4. 一个名为"Data_Interpretation"的说明文件，用于指导数据集的解读与使用。

研究假设：
超声检查已应用于乳用动物乳腺检查领域超过三十年。近年来，已有研究者采用回声纹理分析（echotexture analysis）技术，从乳腺超声图像中提取深层信息，以探究相关病理状态。在我们的研究论文《基于乳腺超声回声纹理分析的深度学习算法预测奶牛日产奶量与生产阶段》中，我们提出如下假设：深度神经网络（deep neural network）可借助乳腺超声图像回声纹理分析得到的特征，实现奶牛日产奶量与生产阶段的预测。

核心研究发现：
深度神经网络（deep neural network）可通过乳腺超声图像的回声纹理分析特征，实现奶牛日产奶量与生产阶段的预测。
本研究的详细发现已刊载于研究论文《基于乳腺超声回声纹理分析的深度学习算法预测奶牛日产奶量与生产阶段》。

数据解读：
文本文件"Data_Interpretation"包含解读与使用本数据集所需的全部信息。关于本数据集采集方法的详细说明，可参见我们的研究论文《基于乳腺超声回声纹理分析的深度学习算法预测奶牛日产奶量与生产阶段》。