LumpySkinDisease_DataHub

Welcome to the Lumpy Skin Disease Data Hub Lumpy Skin Disease (LSD), or nodular skin disease, is a viral disease that mainly affects cattle, particularly cows. It is caused by a poxvirus of the family Poxviridae. The disease gets its name from the nodules or lumps that form on the skin of infected animals. Typical symptoms of LSD include the appearance of nodules on the skin, fever, eye discharge, swelling, and pain in the affected areas. In some cases, the disease can lead to reduced milk production, weight loss and impairment of the animal's general condition. The disease is spread primarily through direct contact between infected animals or through vectors such as mosquitoes and flies. It can have a significant impact on the livestock industry, causing economic losses due to reduced milk production, animal mortality and trade restrictions imposed on affected countries. Prevention of LSD relies mainly on isolation of infected animals, vaccination and control of vectors that transmit the disease. Here you will find how to train a deep learning/machine learning model for LSD-related image classification. This repository is structured using the following folders: code - lumpy_multi_DL_bal.ipynb: This notebook contains code and documentation of a deep learning implementation for image classification for LSD disease. The wording "bal" suggests that the dataset used for training the model is balanced, that is, it contains a similar number of samples for each class. - lumpy_multi_DL_imbal.ipynb: This notebook is similar to the previous one, but uses an unbalanced dataset for model training. This means that some classes may be overrepresented relative to others in the dataset. - lumpy_multi_ML_bal: This notebook contains code and documentation for an implementation of machine learning algorithms for classification of LSD disease images. The wording "bal" suggests that the dataset used is balanced. - lumpy_multi_ML_imbal: Similar to the previous point, but uses an unbalanced dataset for model training. data-processed - lumpy_balanced.zip: This folder contains a set of images related to LSD. The wording "balanced" suggests that the dataset has been balanced, i.e., it contains a similar number of samples for each class of interest, which are necessary for training and evaluation of machine learning or deep learning models. - lumpy_unbalanced.zip: This folder contains a dataset similar to the first, but with the difference that the dataset is unbalanced. This means that some classes of interest may be overrepresented relative to others in the dataset. Within the archive, one may find the same types of data files present in the case of the balanced dataset.欢迎来到结节性皮肤病（Lumpy Skin Disease, LSD）数据中心。结节性皮肤病（LSD）又称结节性皮炎，是一种主要感染牛类（尤其是奶牛）的病毒性疾病，由痘病毒科（Poxviridae）的痘病毒引发。该病因感染动物皮肤上形成的结节或肿块而得名。其典型症状包括皮肤出现结节、发热、眼部分泌物、患处肿胀与疼痛。部分病例可出现产奶量下降、体重减轻以及动物整体健康状况受损的情况。该病主要通过感染动物间的直接接触，或是经由蚊虫、苍蝇等媒介进行传播。该病会对畜牧业造成显著影响，因产奶量降低、动物死亡以及受影响国家实施的贸易限制而带来经济损失。结节性皮肤病的防控主要依赖感染动物隔离、疫苗接种以及疾病传播媒介的管控。在此您将了解如何针对LSD相关图像训练深度学习/机器学习模型。本仓库采用如下文件夹结构： - code 文件夹 - `lumpy_multi_DL_bal.ipynb`：该Jupyter笔记本包含用于LSD图像分类的深度学习实现代码与文档。其中"bal"字样表明训练模型所用的数据集为平衡数据集，即每个类别样本数量相近。 - `lumpy_multi_DL_imbal.ipynb`：该笔记本与上述内容类似，但使用非平衡数据集进行模型训练，即数据集中部分类别的样本占比可能高于其他类别。 - `lumpy_multi_ML_bal`：该笔记本包含用于LSD图像分类的机器学习算法实现代码与文档。其中"bal"字样表明所用数据集为平衡数据集。 - `lumpy_multi_ML_imbal`：与上述内容类似，但使用非平衡数据集进行模型训练。 - data-processed 文件夹 - `lumpy_balanced.zip`：该压缩包包含与LSD相关的图像集。其中"balanced"表明该数据集已完成平衡处理，即每个目标类别拥有相近数量的样本，可用于机器学习或深度学习模型的训练与评估。 - `lumpy_unbalanced.zip`：该压缩包包含与上述平衡数据集类似的数据集，区别在于该数据集为非平衡数据集，即部分目标类别的样本占比可能高于其他类别。压缩包内包含与平衡数据集相同类型的数据文件。