About
This libary is developed for recognizing utility meter digits by using neural network classifier (see the image). The neural network is implemented in Pytorch and the details are available in the code.
The datset (from here + self-made images) has 959 images that are classified into eleven categories: digits "0"-"9" and "10" for cases where digit is not recognizable.
Installation
Install from PyPI:
pip install meter-digits-recognizer
Usage
Minimal example:
import cv2
from meter_digits_recognizer import MeterDigitsRecognizer
# Read image, must be in BGR format (standard in OpenCV)
image = cv2.imread("images/0/1_hot_water_meter_20210212_065239.jpg")
# Run recognizer
dr = MeterDigitsRecognizer()
predictions, confidences = dr.run([image]) # Accepts list of images
# Print
print("Prediction %d (confidence %.1f %%)" % (predictions[0], confidences[0]))
Expected output:
Prediction 0 (confidence 100.0 %)
For additional examples see example.ipynb
Traing
To retrain the neural network follow the steps in the train_neural_net.ipynb notebook.
Credits
- https://github.com/jomjol/neural-network-digital-counter-readout (dataset & inspiration)
1 maintainer
Metadata
1 maintainer
Release files for meter-digits-recognizer 0.2.3
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| meter-digits-recognizer-0.2.3.tar.gz | 2.2 MB | Details |
Release files / meter-digits-recognizer-0.2.3.tar.gz
| Download URL | meter-digits-recognizer-0.2.3.tar.gz |
|---|---|
| Size | 2.2 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
d402541d7885756204e4801eacd21789e2913d67e04bc29c0e04a5eebe592835
|
|
BLAKE2b-256 checksum How to use checksums |
15387843cce5cb2f1e14a3ac07bac12f4751f72390e3290d74bd5806abd805a0
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.1 CPython/3.11.9
|
