Iteration Plan April-May 2021

[innerlee]

This captures the work we planned for the next month, namely, from mid-April to mid-May (5/15). We will improve the user experience including installation, documentation and demonstration. Besides, more algorithms will be supported. Details are

Installation

• Compilation-free: Move cpps to mmcv @jeffreykuang Add pixel group and contour expand ops mmcv#993
• Simplify installation instructions @jeffreykuang simplify installation procedure #188
• Release to pypi @jeffreykuang

Documentation

• Chinese documentation of README @quincylin1 fix pr#70 based on ZZL's comments #77

Demo

• Documentation for demos @quincylin1 Hbsun/demo doc #168 update demo docs #166 add mmocr demo documentation #151
• End2end demo for text detection & recognition @cuhk-hbsun Hbsun/end2end demo #105
• End2end demo of KIE @innerlee
• Colab @quincylin1 MMOCR COlab #147
• Online demo @yuexy
• Add demo cfg @innerlee @cuhk-hbsun

Benchmark

• benchmarks with paddleocr, chineseocr, easyocr, chineseocr_lite
  • Algorithm coverage @jeffreykuang
  • Feature set @jeffreykuang

More Algorithms

• NER @HqWei Ner #106
• FCENet @Zyq-scut add fcenet #133
• DRRG @HolyCrap96 Add drrg #189

---

Deferred

• Online demo @yuexy
• Add demo cfg @innerlee @cuhk-hbsun
• benchmarks with paddleocr, chineseocr, easyocr, chineseocr_lite
  • Algorithm coverage @jeffreykuang
  • Feature set @jeffreykuang
• End2end demo of KIE @innerlee

If you got ideas on what is intersting to implement next, feel free to reply below, or request features here. Happy Research!

[manugarri]

@innerlee are you planing on making single image inference compatible with any model besides DBNET? I have been testing all of the config folder ones and its the only one that doesnt throw an exception.

[yauauau]

Yes but stop sending email to ***@***.***

…

On Thu, Apr 15, 2021, 11:35 AM Manuel Garrido ***@***.***> wrote: @innerlee <https://github.com/innerlee> are you planing on making single image inference compatible with any model besides DBNET? I have been testing all of the config folder ones and its the only one that doesnt throw an exception. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#60 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ATUUQHQHM7JRGY2E2GPS7NDTI4BSHANCNFSM423FJEBA> .

[manugarri]

@yauauau that is not how github works. I havent sent any email, just replied to the GH issue :D

[innerlee]

@manugarri Yes absolutely. The inference should be run with all configs. If you met any problem, don't hesitate to raise an issue :)

[ndcuong91]

Amazing work! I can't wait to test with end2end demo of KIE @innerlee

[manugarri]

@innerlee awesome, does that mean the project is ready to be tested with all configs for inference?

[ndcuong91]

@innerlee Hi bro, what is the status of end2end demo for KIE?

[innerlee]

@ndcuong91 Will work on it next Monday

[seekingdeep]

• Are there plans to implement PAN++?
• What is the current status of the dcn Deformable convolution in DBNet?
• What are the results of DBNet on CTW1500?