# Flowchart of Computer Vision

**URL:** <https://support.prodi.gy/t/flowchart-of-computer-vision/6348>\
**Category:** Uncategorized\
**Tags:** image\
**Created:** [February 7, 2023, 4:28am UTC](https://support.prodi.gy/t/flowchart-of-computer-vision/6348 "2023-02-07T04:28:23Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![kushalrsharma](https://avatars.discourse-cdn.com/v4/letter/k/b9e5f3/32.png) [@kushalrsharma](https://support.prodi.gy/u/kushalrsharma)\
**Post date:** [February 7, 2023, 4:28am UTC](https://support.prodi.gy/t/flowchart-of-computer-vision/6348/1 "2023-02-07T04:28:23Z")

</div>

Can I have the flowchart for the computer Vision?

---

<div class="post-metadata">

**Author:** ![ryanwesslen](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/ryanwesslen/32/2969_2.png) [@ryanwesslen](https://support.prodi.gy/u/ryanwesslen)\
**Post date:** [February 7, 2023, 12:51pm UTC](https://support.prodi.gy/t/flowchart-of-computer-vision/6348/2 "2023-02-07T12:51:21Z")

</div>

hi @kushalrsharma!

> [@kushalrsharma](#):
>
> Can I have the flowchart for the computer Vision?

Thanks for your question. The challenge is that we (Explosion) don't natively have a CV specific package for training. Prodigy offers CV annotation but for training you'd need to use some some third-party framework like [TensorFlow](https://prodi.gy/docs/computer-vision#tensorflow) or [PyTorch](https://prodi.gy/docs/custom-recipes#video-imagecaptioning). Flowcharts are about making recommendations and the problem is we're not able to make specific recommendations on third-party tools as we leave that to users on what works best for them.

---

<div class="post-metadata">

**Author:** ![kushalrsharma](https://avatars.discourse-cdn.com/v4/letter/k/b9e5f3/32.png) [@kushalrsharma](https://support.prodi.gy/u/kushalrsharma)\
**Post date:** [February 8, 2023, 2:55am UTC](https://support.prodi.gy/t/flowchart-of-computer-vision/6348/4 "2023-02-08T02:55:31Z")

</div>

This means i need to first annotate using prodigy and need to make a script to convert them to either YOLO format or COCO format for training. Or is there any scripts that helps me to do so?

---

<div class="post-metadata">

**Author:** ![ryanwesslen](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/ryanwesslen/32/2969_2.png) [@ryanwesslen](https://support.prodi.gy/u/ryanwesslen)\
**Post date:** [February 8, 2023, 12:17pm UTC](https://support.prodi.gy/t/flowchart-of-computer-vision/6348/5 "2023-02-08T12:17:40Z")

</div>

> [@kushalrsharma](#):
>
> This means i need to first annotate using prodigy and need to make a script to convert them to either YOLO format or COCO format for training. Or is there any scripts that helps me to do so?

Check out these two posts:

> [@image classification output for Yolo in 2022](https://support.prodi.gy/t/image-classification-output-for-yolo-in-2022/5156/2):
>
> Hi @info2000, the output annotations are still in a JSONL file, and you still need to parse it to the YOLO format afterwards. You can refer to this [blogpost](https://mlops.systems/redactionmodel/computervision/datalabelling/2021/11/29/prodigy-object-detection-training.html) that did a similar approach (although here he's converting it into the COCO format but he also showed what a typical image annotation looks like)

> [@Datasets and using pre-annotated data](https://support.prodi.gy/t/datasets-and-using-pre-annotated-data/1476/26):
>
> I just made a demo COCO display recipe that allows me to simply visualize the dataset and review it for correctness It might be a good starting point for your own workflow

If you find something that works, we would be grateful if you could post to help others like you in the future!