search

RapidAI / LabelConvert

🔄 A tool for object detection and image segmentation dataset format conversion.
https://rapidai.github.io/LabelConvert/docs
Apache License 2.0
292 stars 67 forks source link

語意分割的YOLO如何轉COCO #6

Closed stphtan94117 closed 7 months ago

stphtan94117 commented 1 year ago

作者你好 由於yolov5跟yolov7都有出了語意分割(segmentation)，因此save_txt的內容，就不像傳統的只有bbox xywh值 而是有多個點組成的polygon

不曉得作者您能否做一個polygon的yolo(txt)轉COCO (json)方式 謝謝

SWHL commented 1 year ago
stphtan94117 commented 1 year ago

你好 這方面我也不是寫CODE高手，我只能提供相關yolo segmentation 檔案給你參考 讓你打開TXT檔，看它內容長怎樣 https://universe.roboflow.com/chonnam-national-university/crack-bphdr/dataset/1

下載時候，麻煩選擇yolov5 or v7 pytorch，裡面就是txt了

SWHL commented 1 year ago

嗯嗯，你这个需求作为TODO，有空我看一下

stphtan94117 commented 1 year ago

謝謝你 再麻煩妳新增yolo2coco、coco2yolo 或者正棒的是 json2yolo互轉

bit-scientist commented 1 year ago

Well, here we are. @SWHL did you happen to make it?

SWHL commented 1 year ago

I want to do this but don't have the time. Your contributions are welcome, if you can.

Youho99 commented 7 months ago

The YoloSeg2COCO functionality also interests me. Any news on a possible implementation?

SWHL commented 7 months ago

It's time to start solving the issue.

SWHL commented 7 months ago

在label_convert >=0.0.10中，已经支持了yolo seg格式转为coco格式。label_conert会自动处理seg类标注。具体可参见文档：link

In label_convert >=0.0.10, the conversion of yolo seg format to coco format is already supported. label_conert will automatically handle seg class annotations. For details, please refer to the document: link

SWHL commented 7 months ago

This issue will be closed. If you have any question, welcome to reopen it.