https://github.com/simon2k/pneumonia-image-classifier
Predict if a given photo represents lungs with pneumonia.
https://github.com/simon2k/pneumonia-image-classifier
classfication computer-vision image-classification machine-learning pneumonia-detection tensorflow
Last synced: 4 months ago
JSON representation
Predict if a given photo represents lungs with pneumonia.
- Host: GitHub
- URL: https://github.com/simon2k/pneumonia-image-classifier
- Owner: simon2k
- Created: 2019-09-22T17:58:08.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2019-09-23T23:56:28.000Z (almost 6 years ago)
- Last Synced: 2025-01-12T18:25:01.741Z (6 months ago)
- Topics: classfication, computer-vision, image-classification, machine-learning, pneumonia-detection, tensorflow
- Language: Jupyter Notebook
- Homepage: https://github.com/simon2k/pneumonia-image-classifier/blob/master/pneumonia-detection-softmax.ipynb
- Size: 7.81 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Pneumonia Image Classifier
The intention is to predict if a given photo represents lungs with pneumonia.
Dataset: https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia
## Model Summary
The final model was trained for about 25 epochs.
Scores for the validation set:
- Loss: 0.44
- Accuracy: 75.0%
- Precision score: 30.0%
- Recall score: 37.5%
- F1 score: 33.33%
Confusion matrix:
## Model Architecture
The model is based on InceptionV3 model with two fully connected top layers:
* `x = Flatten()(x)`
* `x = Dense(64, activation='relu')(x)`
* `x = Dense(2, activation='softmax')(x)`
Summary:
* Total params: 30,191,586
* Trainable params: 8,388,802
* Non-trainable params: 21,802,784