Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

recognition rate is very low #37

Open
lbbboy opened this issue Jul 9, 2022 · 4 comments
Open

recognition rate is very low #37

lbbboy opened this issue Jul 9, 2022 · 4 comments

Comments

@lbbboy
Copy link

lbbboy commented Jul 9, 2022

Device

  • For PC, what is your screen resolution: 1080P
  • For PC, what is your OS: win11
  • Game language:Simplified Chinese traditional Chinese
  • App version: V1.7

** Problem(s) **
Please describe the issue in as much detail.
Almost unusable, only 9 out of 50 images are recognized, and they are all wrong
微信图片_20220710072339

{'slots': [3, 0, 0], 'skills': {}, 'errors': [['霍 層 作 化', '1', 'ParseError.MUST_FIX'], ['城 垂 跖 離 UP', '2', 'ParseError.MUST_FIX']]}
frames\frame0.png
frames\frame196.png
frames\frame208.png
frames\frame253.png
frames\frame474.png
frames\frame486.png

{'slots': [3, 1, 0], 'skills': {}, 'errors': [['不 聞', '1', 'ParseError.MUST_FIX'], ['籟 酎 怯', '2', 'ParseError.MUST_FIX']]}
frames\frame112.png
frames\frame148.png
frames\frame220.png
frames\frame522.png

{'slots': [2, 1, 0], 'skills': {}, 'errors': [['泥 霈 耐 作', '1', 'ParseError.MUST_FIX'], ['披 分 術 [ 力 ]', '1', 'ParseError.MUST_FIX']]}
frames\frame122.png
frames\frame134.png
frames\frame172.png
frames\frame183.png
frames\frame242.png
frames\frame278.png
frames\frame289.png
frames\frame312.png
frames\frame346.png
frames\frame357.png
frames\frame404.png
frames\frame439.png
frames\frame451.png
frames\frame99.png

{'slots': [1, 1, 0], 'skills': {}, 'errors': [['砲 遼 褲', '1', 'ParseError.MUST_FIX']]}
frames\frame14.png
frames\frame231.png
frames\frame266.png
frames\frame381.png
frames\frame39.png
frames\frame416.png
frames\frame499.png
frames\frame63.png
frames\frame75.png

{'slots': [1, 0, 0], 'skills': {}, 'errors': [['L E', '1', 'ParseError.MUST_FIX'], ['怎 惇', '2', 'ParseError.MUST_FIX']]}
frames\frame160.png
frames\frame28.png
frames\frame301.png
frames\frame324.png
frames\frame511.png
frames\frame547.png
frames\frame559.png

{'slots': [1, 1, 1], 'skills': {}, 'errors': [['居 佑 賭 帚 跋 慈 的 酒', '1', 'ParseError.MUST_FIX'], ['張 化 括 纏', '1', 'ParseError.MUST_FIX']]}
frames\frame335.png
frames\frame427.png
frames\frame535.png
frames\frame87.png

{'slots': [3, 2, 0], 'skills': {}, 'errors': [['耳 壑', '1', 'ParseError.MUST_FIX'], ['泡 沬 2 舞', '1', 'ParseError.MUST_FIX']]}
frames\frame369.png
frames\frame50.png

{'slots': [0, 0, 0], 'skills': {}, 'errors': [['力 星 鯨', '1', 'ParseError.MUST_FIX'], ['粟 降 寺 孵 張', '1', 'ParseError.MUST_FIX']]}
frames\frame463.png
frames\frame571.png
frames\frame584.png

video

@lbbboy lbbboy added the other label Jul 9, 2022
@chpoit
Copy link
Owner

chpoit commented Jul 10, 2022

As much as I want to help you, languages with kanji are very hard to parse for tesseract, espescially considering the resolution of the switch recordings. The amount of detail in a single character is quite high and as good as Google's OCR data might be, it is not perfect.

I can try to improve this, but I don't think I will be able to fully resolve the issue, you just have to look at the second character in the first frame of your video, it's already blurry for humans, a computer will struggle even more.
Screenshot_20220710_090519

Thank you for your video, I can at least use it to attempt to improve detection.

The "easiest" solution would be for you to add automatic corrections in the corrections.chi_tra.csv or corrections.chi_sim.csv in your Appdata/Local/utsushis-charm directory.

As an aside, I am curious, you seem to understand english well, why don't you record in english and use english set builders? Are you simply too used to the chinese names for everything?

@chpoit chpoit added the kanji label Jul 10, 2022
@lbbboy
Copy link
Author

lbbboy commented Jul 12, 2022

I found an easier way

MHR charm json to txt converter

@Vessile822
Copy link

src.charm_extraction - WARNING - Skill-less charm found in frames
Why cant recognize?

@chpoit
Copy link
Owner

chpoit commented Aug 19, 2022

src.charm_extraction - WARNING - Skill-less charm found in frames Why cant recognize?

I am afraid I cannot help you with this little information. Either create a complete issue, or submit proper information related to this issue. Thank you.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

3 participants