https://github.com/pjarbas/image-classification

Image classification with tensorflow/keras
https://github.com/pjarbas/image-classification

computer-vision deep-learning keras opencv tensorboard tensorflow

Last synced: 2 months ago
JSON representation

Image classification with tensorflow/keras

• Host: GitHub
• URL: https://github.com/pjarbas/image-classification
• Owner: PJarbas
• License: mit
• Created: 2022-08-12T13:33:44.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2022-09-16T13:38:41.000Z (almost 4 years ago)
• Last Synced: 2025-03-14T11:31:55.377Z (over 1 year ago)
• Topics: computer-vision, deep-learning, keras, opencv, tensorboard, tensorflow
• Language: Python
• Homepage:
• Size: 205 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Image-classification 

This project aims to use image classification using transfer learning with pre-trained models: VGG19, ResNet50, InceptionV3.

### Get started

Ensure that your **python** version is >= 3.9 

Train the model using the following command:

```bash

   $ cd src

   $ python train.py 

```

### Results 

* The pre-trained models seem to overfitting, and the baseline model for the cifar10 dataset can be enough.

* We can do data augmentation to improve the results.

Accuracy                              |  Loss                             

:------------------------------------:|:---------------------------------:

![](results/baseline-accuracy.png)    | ![](results/baseline-loss.png) 

:------------------------------------:|----------------------------------:

![](results/resnet50-accuracy.png)    | ![](results/resnet50-loss.png) 

:------------------------------------:|----------------------------------:

![](results/mobilenet_v2-accuracy.png)| ![](results/mobilenet_v2-loss.png) 

### Classify the images

run the main script using the following command:

```bash

   $ cd src

   $ python main.py 

```

### Tensorboard

```bash

   $ tensorboard --logdir logs

```