chesslablab / scanblab Goto Github PK
View Code? Open in Web Editor NEWChess piece image recognizer.
License: MIT License
scanblab's People
Contributors
Stargazers
scanblab's Issues
Visual recognition of chess pieces
It seems as if this problem should be quite similar to training a model that recognizes the numbers in handwritten digits.
See:
Keep it up, and happy coding and learing!
Recognize empty squares
See Chess\Media\JpgToPgnAn in PHP Chess.
Use more samples to train the multilayer neural network
The process of training the multilayer neural network may require some trial and error until the final result is achieved. At this moment the results are perhaps not too accurate recognizing knights.
{
"breakdown": {
"overall": {
"accuracy": 0.9994216990880235,
"balanced accuracy": 0.9969842248240661,
"f1 score": 0.9941602196186778,
"precision": 0.9940627491890405,
"recall": 0.9942795094934874,
"specificity": 0.9996889401546447,
"negative predictive value": 0.9996881232795347,
"false discovery rate": 0.005937250810959638,
"miss rate": 0.005720490506512628,
"fall out": 0.00031105984535529675,
"false omission rate": 0.0003118767204653641,
"mcc": 0.9938551321222888,
"informedness": 0.9939684496481321,
"markedness": 0.993750872468575,
"true positives": 19924,
"true negatives": 239088,
"false positives": 75,
"false negatives": 75,
"cardinality": 19999
},
"classes": {
"1": {
"accuracy": 0.9987468043510953,
"balanced accuracy": 0.9814868703344161,
"f1 score": 0.962178517397882,
"precision": 0.9607250755287009,
"recall": 0.9636363636363636,
"specificity": 0.9993373770324685,
"negative predictive value": 0.9993883168518707,
"false discovery rate": 0.0392749244712991,
"miss rate": 0.036363636363636376,
"fall out": 0.000662622967531501,
"false omission rate": 0.0006116831481293161,
"informedness": 0.9629737406688321,
"markedness": 0.9601133923805716,
"mcc": 0.9615425029227576,
"true positives": 318,
"true negatives": 19606,
"false positives": 13,
"false negatives": 12,
"cardinality": 330,
"proportion": 0.01650082504125206
},
"N": {
"accuracy": 0.997546688028839,
"balanced accuracy": 0.9941865674490369,
"f1 score": 0.9850290253589978,
"precision": 0.9799392097264438,
"recall": 0.9901719901719902,
"specificity": 0.9982011447260835,
"negative predictive value": 0.9991270187690965,
"false discovery rate": 0.020060790273556228,
"miss rate": 0.009828009828009798,
"fall out": 0.0017988552739165442,
"false omission rate": 0.0008729812309035134,
"informedness": 0.9883731348980738,
"markedness": 0.9790662284955403,
"mcc": 0.9837086751325164,
"true positives": 1612,
"true negatives": 18312,
"false positives": 33,
"false negatives": 16,
"cardinality": 1628,
"proportion": 0.08140407020351018
},
...
"r": {
"accuracy": 1,
"balanced accuracy": 1,
"f1 score": 1,
"precision": 1,
"recall": 1,
"specificity": 1,
"negative predictive value": 1,
"false discovery rate": 0,
"miss rate": 0,
"fall out": 0,
"false omission rate": 0,
"informedness": 1,
"markedness": 1,
"mcc": 1,
"true positives": 1622,
"true negatives": 18302,
"false positives": 0,
"false negatives": 0,
"cardinality": 1622,
"proportion": 0.08110405520276014
}
}
},
"matrix": {
"1": {
"1": 318,
"N": 0,
"B": 8,
"Q": 0,
"b": 0,
"K": 0,
"k": 0,
"R": 0,
"n": 0,
"P": 5,
"p": 0,
"q": 0,
"r": 0
},
"N": {
"1": 12,
"N": 1612,
"B": 7,
"Q": 0,
"b": 0,
"K": 0,
"k": 0,
"R": 0,
"n": 0,
"P": 14,
"p": 0,
"q": 0,
"r": 0
},
...
"r": {
"1": 0,
"N": 0,
"B": 0,
"Q": 0,
"b": 0,
"K": 0,
"k": 0,
"R": 0,
"n": 0,
"P": 0,
"p": 0,
"q": 0,
"r": 1622
}
}
}
It'd be convenient to use 50,000 samples to train the multilayer neural network rather than 20,000 and compare the results obtained with the current ones. If the results aren't significantly improved, then the sample data may require some tweaking.
This is a few step process as described in the README file.
- Remove all files in the testing and the training folders.
- Prepare 50,000 samples for further training.
- Train the neural network.
- Prepare 50,000 samples for further testing.
- Make predictions.
Keep it up, and happy learning and coding.
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.