# Pose estimation after object detection or start over?

**URL:** <https://discuss.roboflow.com/t/pose-estimation-after-object-detection-or-start-over/5205>\
**Category:** 🤝  Community Help\
**Created:** [April 8, 2024, 4:55pm UTC](https://discuss.roboflow.com/t/pose-estimation-after-object-detection-or-start-over/5205 "2024-04-08T16:55:46Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![wschaft](https://avatars.discourse-cdn.com/v4/letter/w/67e7ee/32.png) [@wschaft](https://discuss.roboflow.com/u/wschaft)\
**Post date:** [April 8, 2024, 4:55pm UTC](https://discuss.roboflow.com/t/pose-estimation-after-object-detection-or-start-over/5205/1 "2024-04-08T16:55:46Z")

</div>

So I’ve created a custom dataset which I’ve trained for object detection. It consists of a few classes, with some of them are “male” or “female” players. I can use it for tracking players in a sports video.

However, now I want to do something with pose estimation. Like if a player shoots or passes a ball. Should I create a new project and dataset for keypoint detection? Or can I use my current object detection dataset and make use of the detected bounding boxes and doing pose estimation per bounding box?

Just looking for the right approach. Thanks in advance!

---

<div class="post-metadata">

**Author:** ![system](https://canada1.discourse-cdn.com/flex029/uploads/roboflow1/original/1X/bb6edbfa56d0cf2ee603f061e9e09103502cc8cc.png) [@system](https://discuss.roboflow.com/u/system)\
**Post date:** [April 29, 2024, 4:56pm UTC](https://discuss.roboflow.com/t/pose-estimation-after-object-detection-or-start-over/5205/2 "2024-04-29T16:56:41Z")

</div>

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.