mmpose----AnimalKingdom数据集使用mmpose训练

本文主要讲述了使用mmpose训练AnimalKingdom数据集来预测鸟类骨架

1、下载mmpose，AnimalKingdom数据集补充代码

安装 — MMPose 1.0.0 文档根据官方文档安装mmpose可以跑通demo

GitHub - sutdcv/Animal-Kingdom: [CVPR2022] Animal Kingdom: A Large and Diverse Dataset for Animal Behavior Understanding 

2、将AK补充代码下的ak.py复制到到mmpose-main的configs\ _base_\datasets 

 3、在mmpose\datasets\datasets\animal\__init__.py里注册AK数据集,在头文件from .ak_dataset import AKDataset，并且在此文件下创建一个ak_dataset.py,将下边的内容复制进去

新建一个文件将代码复制进去，注册AK数据集

from ..base import BaseCocoStyleDataset
@DATASETS.register_module()
class AKDataset(BaseCocoStyleDataset):

    """Animal Kingdom dataset for animal pose estimation.

    "[CVPR2022] Animal Kingdom: A Large and Diverse Dataset for Animal Behavior Understanding"
    More details can be found in the `paper <https://www.researchgate.net/publication/359816954_Animal_Kingdom_A_Large_and_Diverse_Dataset_for_Animal_Behavior_Understanding>`__ .

    Website: <https://sutdcv.github.io/Animal-Kingdom>
        
    The dataset loads raw features and apply specified transforms
    to return a dict containing the image tensors and other information.

    Animal Kingdom keypoint indexes::

        0: 'Head_Mid_Top',
        1: 'Eye_Left',
        2: 'Eye_Right',
        3: 'Mouth_Front_Top',
        4: 'Mouth_Back_Left',
        5: 'Mouth_Back_Right',
        6: 'Mouth_Front_Bottom',
        7: 'Shoulder_Left',
        8: 'Shoulder_Right',
        9: 'Elbow_Left',
        10: 'Elbow_Right',
        11: 'Wrist_Left',
        12: 'Wrist_Right',
        13: 'Torso_Mid_Back',
        14: 'Hip_Left',
        15: 'Hip_Right',
        16: 'Knee_Left',
        17: 'Knee_Right',
        18: 'Ankle_Left ',
        19: 'Ankle_Right',
        20: 'Tail_Top_Back',
        21: 'Tail_Mid_Back',
        22: 'Tail_End_Back

    Args:
        ann_file (str): Annotation file path. Default: ''.
        bbox_file (str, optional): Detection result file path. If
            ``bbox_file`` is set, detected bboxes loaded from this file will
            be used instead of ground-truth bboxes. This setting is only for
            evaluation, i.e., ignored when ``test_mode`` is ``False``.
            Default: ``None``.
        data_mode (str): Specifies the mode of data samples: ``'topdown'`` or
            ``'bottomup'``. In ``'topdown'`` mode, each data sample contains
            one instance; while in ``'bottomup'`` mode, each data sample
            contains all instances in a image. Default: ``'topdown'``
        metainfo (dict, optional): Meta information for dataset, such as class
            information. Default: ``None``.
        data_root (str, optional): The root directory for ``data_prefix`` and
            ``ann_file``. Default: ``None``.
        data_prefix (dict, optional): Prefix for training data. Default:
            ``dict(img=None, ann=None)``.
        filter_cfg (dict, optional): Config for filter data. Default: `None`.
        indices (int or Sequence[int], optional): Support using first few
            data in annotation file to facilitate training/testing on a smaller
            dataset. Default: ``None`` which means using all ``data_infos``.
        serialize_data (bool, optional): Whether to hold memory using
            serialized objects, when enabled, data loader workers can use
            shared RAM from master process instead of making a copy.
            Default: ``True``.
        pipeline (list, optional): Processing pipeline. Default: [].
        test_mode (bool, optional): ``test_mode=True`` means in test phase.
            Default: ``False``.
        lazy_init (bool, optional): Whether to load annotation during
            instantiation. In some cases, such as visualization, only the meta
            information of the dataset is needed, which is not necessary to
            load annotation file. ``Basedataset`` can skip load annotations to
            save time by set ``lazy_init=False``. Default: ``False``.
        max_refetch (int, optional): If ``Basedataset.prepare_data`` get a
            None img. The maximum extra number of cycles to get a valid
            image. Default: 1000.
    """

    METAINFO: dict = dict(from_file='configs/_base_/datasets/ak.py')

4、仿照着写一个配置文件 

mmpose-main\configs\animal_2d_keypoint\topdown_heatmap\animalpose随便挑一个仿着写

 首先更改hook，改为pck准确率计算

接着更改模型的输出，AK的输出骨架时23个节点，因此把out_channels更改为23

更改数据集的名字，和图片和标签的路径

更改标签路径，并在img填上图片文件夹的名字

还可以改batch和其他参数

最后将val也改为PCK，给文件改个名

 

5、开始训练

直接终端训练，找到刚才的配置文件

python tools/train.py configs/animal_2d_keypoint/topdown_heatmap/ak/td-hm_hrnet-w32_8xb64-210e_ak-256x256.py

 6、得到测试结果

官方有出demo可以预测视频图片Demos — MMPose 1.0.0 文档

预测需要同时全部安装mmdet和mmengine,按照官方文档，全部安装即可

###安装mmdet
git clone https://github.com/open-mmlab/mmdetection.git
cd mmdetection
pip install -v -e .

#安装mmengine
git clone https://github.com/open-mmlab/mmengine.git
cd mmengine
pip install -e . -v

安装完成后，我是预测鸟，所以id为14，检测的模型和py文件运行官方的demo会下载

!python demo/topdown_demo_with_mmdet.py \
    demo/mmdetection_cfg/faster_rcnn_r50_fpn_coco.py \
    faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth \
    configs/animal_2d_keypoint/topdown_heatmap/ak/td-hm_hrnet-w32_1xb32-210e_ak-256x256.py \
    best_PCK_epoch_210.pth \
    --input new.mp4 \
    --output-root vis_result --det-cat-id=14

最终效果还不错能识别多只鸟并且加上骨架