type = 'MultiViewMultiPersonTopDownEstimator'
bbox_thr = 0.9
work_dir = './temp'
verbose = False
logger = None
pred_kps3d_convention = 'coco'

bbox_detector = dict(
    type='MMdetDetector',
    mmdet_kwargs=dict(
        checkpoint='weight/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth',
        config='configs/modules/human_perception/' +
        'mmdet_faster_rcnn_r50_fpn_coco.py',
        device='cuda'),
    batch_size=10)

kps2d_estimator = dict(
    type='MMposeTopDownEstimator',
    mmpose_kwargs=dict(
        checkpoint='weight/hrnet_w48_coco_wholebody' +
        '_384x288_dark-f5726563_20200918.pth',
        config='configs/modules/human_perception/mmpose_hrnet_w48_' +
        'coco_wholebody_384x288_dark_plus.py',
        device='cuda'),
    bbox_thr=bbox_thr)

associator = dict(
    type='MvposeAssociator',
    triangulator=dict(
        type='AniposelibTriangulator',
        camera_parameters=[],
        logger=logger,
    ),
    affinity_estimator=dict(type='AppearanceAffinityEstimator', init_cfg=None),
    point_selector=dict(
        type='HybridKps2dSelector',
        triangulator=dict(
            type='AniposelibTriangulator', camera_parameters=[],
            logger=logger),
        verbose=verbose,
        ignore_kps_name=['left_eye', 'right_eye', 'left_ear', 'right_ear'],
        convention=pred_kps3d_convention),
    multi_way_matching=dict(
        type='MultiWayMatching',
        use_dual_stochastic_SVT=True,
        lambda_SVT=50,
        alpha_SVT=0.5,
        n_cam_min=3,
    ),
    kalman_tracking=dict(type='KalmanTracking', n_cam_min=3, logger=logger),
    identity_tracking=dict(
        type='KeypointsDistanceTracking',
        tracking_distance=1,
        tracking_kps3d_convention=pred_kps3d_convention,
        tracking_kps3d_name=[
            'left_shoulder', 'right_shoulder', 'left_hip_extra',
            'right_hip_extra'
        ]),
    checkpoint_path='./weight/mvpose/' +
    'resnet50_reid_camstyle-98d61e41_20220921.pth',
    best_distance=600,
    interval=5,
    bbox_thr=bbox_thr,
    device='cuda',
    logger=logger,
)

smplify = dict(
    type='SMPLify',
    verbose=True,
    info_level='stage',
    n_epochs=1,
    use_one_betas_per_video=False,
    hooks=[
        dict(type='SMPLifyVerboseHook'),
    ],
    logger=logger,
    body_model=dict(
        type='SMPL',
        gender='neutral',
        num_betas=10,
        keypoint_convention='smpl_45',
        model_path='xrmocap_data/body_models/smpl',
        batch_size=1,
        logger=logger),
    optimizer=dict(
        type='LBFGS', max_iter=20, lr=1.0, line_search_fn='strong_wolfe'),
    ignore_keypoints=[
        'neck_openpose', 'right_hip_openpose', 'left_hip_openpose',
        'right_hip_extra', 'left_hip_extra'
    ],
    handlers=[
        dict(
            handler_key='keypoints3d_mse',
            type='Keypoint3dMSEHandler',
            mse_loss=dict(
                type='KeypointMSELoss',
                loss_weight=10.0,
                reduction='sum',
                sigma=100),
            logger=logger),
        dict(
            handler_key='shape_prior',
            type='BetasPriorHandler',
            prior_loss=dict(
                type='ShapePriorLoss', loss_weight=5e-3, reduction='mean'),
            logger=logger),
        dict(
            handler_key='joint_prior',
            type='BodyPosePriorHandler',
            prior_loss=dict(
                type='JointPriorLoss',
                loss_weight=20.0,
                reduction='sum',
                smooth_spine=True,
                smooth_spine_loss_weight=20,
                use_full_body=True),
            logger=logger),
        dict(
            handler_key='pose_prior',
            type='BodyPosePriorHandler',
            prior_loss=dict(
                type='MaxMixturePriorLoss',
                prior_folder='xrmocap_data/body_models',
                num_gaussians=8,
                loss_weight=4.78**2,
                reduction='sum'),
            logger=logger),
        dict(
            handler_key='pose_reg',
            type='BodyPosePriorHandler',
            prior_loss=dict(
                type='PoseRegLoss', loss_weight=0.001, reduction='mean'),
            logger=logger),
        dict(
            handler_key='keypoints3d_limb_len',
            type='Keypoint3dLimbLenHandler',
            loss=dict(
                type='LimbLengthLoss',
                convention='smpl',
                loss_weight=1.0,
                reduction='mean'),
            logger=logger),
    ],
    stages=[
        # stage 0
        dict(
            n_iter=10,
            ftol=1e-4,
            fit_global_orient=False,
            fit_transl=False,
            fit_body_pose=False,
            fit_betas=True,
            keypoints3d_mse_weight=0.0,
            keypoints3d_limb_len_weight=1.0,
            shape_prior_weight=5e-3,
            joint_prior_weight=0.0,
            pose_reg_weight=0.0,
            pose_prior_weight=0.0),
        # stage 1
        dict(
            n_iter=50,
            ftol=1e-4,
            fit_global_orient=True,
            fit_transl=True,
            fit_body_pose=False,
            fit_betas=False,
            keypoints3d_mse_weight=1.0,
            keypoints3d_limb_len_weight=0.0,
            shape_prior_weight=0.0,
            joint_prior_weight=0.0,
            pose_reg_weight=0.0,
            pose_prior_weight=0.0,
            body_weight=0.0,
            face_weight=0.0,
            hand_weight=0.0,
            shoulder_weight=5.0,
            hip_weight=5.0,
            use_shoulder_hip_only=True),
        # stage 2
        dict(
            n_iter=120,
            ftol=1e-4,
            fit_global_orient=True,
            fit_transl=True,
            fit_body_pose=True,
            fit_betas=False,
            keypoints3d_mse_weight=10.0,
            keypoints3d_limb_len_weight=0.0,
            shape_prior_weight=0.0,
            joint_prior_weight=0.0,
            pose_reg_weight=0.001,
            pose_prior_weight=0.0,
            body_weight=1.0,
            use_shoulder_hip_only=False),
    ],
)

triangulator = dict(
    type='AniposelibTriangulator',
    camera_parameters=[],
    logger=logger,
)
point_selectors = [
    dict(
        type='ReprojectionErrorPointSelector',
        target_camera_number=2,
        triangulator=dict(
            type='AniposelibTriangulator', camera_parameters=[],
            logger=logger),
        verbose=verbose,
        logger=logger,
    )
]

kps3d_optimizers = [
    dict(type='TrajectoryOptimizer', verbose=verbose, logger=logger),
    dict(type='NanInterpolation', verbose=verbose, logger=logger),
    # SMPLShapeAwareOptimizer is optional.
    dict(
        type='SMPLShapeAwareOptimizer',
        smplify=smplify,
        body_model=smplify['body_model'],
        projector=dict(type='PytorchProjector', camera_parameters=[]),
        iteration=1,
        refine_threshold=1,
        kps2d_conf_threshold=0.97,
        use_percep2d_optimizer=False,
        verbose=verbose,
        logger=logger),
    # After SMPL shape-aware optimizer, the keypoints are not very stable,
    # so trajectory optimization is added.
    dict(type='TrajectoryOptimizer', verbose=verbose, logger=logger),
    dict(type='NanInterpolation', verbose=verbose, logger=logger),
]