Object Detection toolkit based on PaddlePaddle. It supports object detection, instance segmentation, multiple object tracking and real-time multi-person keypoint detection.
问题:
1.我自己标注的关键点名称和coco的不一样,可以不一样吗?
2.自己标注的关键点顺序可以和coco的不一样吗?
3.我标注了4个关键点,为啥会出现IndexError: index 4 is out of bounds for axis 0 with size 4问题?
教程中coco的如下:(教程https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.3/docs/tutorials/PrepareKeypointDataSet_cn.md)
问题: 1.我自己标注的关键点名称和coco的不一样,可以不一样吗? 2.自己标注的关键点顺序可以和coco的不一样吗? 3.我标注了4个关键点,为啥会出现IndexError: index 4 is out of bounds for axis 0 with size 4问题? 教程中coco的如下:(教程https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.3/docs/tutorials/PrepareKeypointDataSet_cn.md)![image](https://github.com/PaddlePaddle/PaddleDetection/assets/49717982/c1a80bc6-5407-422d-aa99-68b396b29d79)
关键点信息如下: categories = [ { "supercategory": "person", "id": 1, "name": "person", "keypoints": [ "l_hand", "r_hand", "l_elbow", "r_elbow" ], "skeleton": [ [1, 3], [2, 4] ] } ] 依据dark_hrnet_w32_256x192.yml文件配置模型,该yml文件上半部分内容如下: use_gpu: true log_iter: 5 save_dir: output snapshot_epoch: 10 weights: output/hrnet_w32_256x192/model_final_pick_door epoch: 210
num_joints: &num_joints 17
num_joints: &num_joints 4 pixel_std: &pixel_std 200 metric: KeyPointTopDownCOCOEval
num_classes: 1
num_classes: 4 train_height: &train_height 256 train_width: &train_width 192 trainsize: &trainsize [train_width, train_height] hmsize: &hmsize [48, 64]
flip_perm: &flip_perm [[1, 2], [3, 4], [5, 6], [7, 8], [9, 10], [11, 12], [13, 14], [15, 16]]
flip_perm: &flip_perm [[1, 2], [3, 4]]