Image recognition based on deep learning in Haemonchus contortus motility assays

The repository contains the data associated with the paper `Image recognition based on deep learning in Haemonchus contortus motility assays`. The following folders form part of the repository: - Annotation Data - annotated microscope images used for the training of the Mask R-CNN model. The data are divided into `train` and `val` - Mask R-CNN - contains the trained weights for the Mask R-CNN model - Motility Output - output of the 3 compared algorithms (Wiggle Index, WF-NTP and Mask R-CNN). - Motility Videos - input videos used for the motility detection. The naming convetno is `XXXzYYY.avi`, where `XXX` denotes the motility group and `YYY` the sequence number for the video within a given group

本仓库包含与论文《基于深度学习的捻转血矛线虫（Haemonchus contortus）运动试验图像识别》相关的配套数据。本仓库包含以下文件夹：
- 标注数据（Annotation Data）：用于训练Mask R-CNN模型的标注显微镜图像，数据划分为训练集（`train`）与验证集（`val`）。
- Mask R-CNN：存放该模型的训练权重。
- 运动分析输出（Motility Output）：3种对比算法（摆动指数（Wiggle Index）、WF-NTP及Mask R-CNN）的输出结果。
- 运动检测视频（Motility Videos）：用于运动检测的原始输入视频，命名规则为`XXXzYYY.avi`，其中`XXX`代表运动组别，`YYY`为该组别内视频的序列号。