https://github.com/praneetneuro/convolutional-autoencoder-for-image-transformation

Can be used to build autoencoder models for image transformation like image colorization
https://github.com/praneetneuro/convolutional-autoencoder-for-image-transformation

Last synced: 7 days ago
JSON representation

Can be used to build autoencoder models for image transformation like image colorization

• Host: GitHub
• URL: https://github.com/praneetneuro/convolutional-autoencoder-for-image-transformation
• Owner: PraneetNeuro
• License: mit
• Created: 2021-01-10T13:42:04.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2021-01-18T07:34:52.000Z (almost 4 years ago)
• Last Synced: 2024-05-14T00:22:59.613Z (6 months ago)
• Language: Python
• Size: 8.79 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Convolutional-Autoencoder-for-Image-transformation

Can be used to build autoencoder models for image transformation like image colorization

# Note:

Feel free to tinker with the source file to work with images of different dimensions

## Sample code for training the model and inference

```python

dataset = Dataset('SRC_IMAGES_PATH', 'TARGET_IMAGES_PATH')

autoencoder = AutoEncoder(dataset, epochs=35)

def inference(images_path, save_path, ground_truth_path=None):

    if ground_truth_path is not None:

        for img_n in tqdm(os.listdir(images_path)):

            try:

                img = cv2.imread(images_path + img_n)

                img_ = np.array(cv2.resize(img, (100, 100))) / 255

                img = np.expand_dims(img_, 0)

                output = np.array(autoencoder.model.predict([np.array(img)])[0])

                target = cv2.imread(ground_truth_path + img_n)

                target = np.array(cv2.resize(target, (100, 100)))

                res = np.concatenate((img_ * 255, output * 255, target), axis=1)

                cv2.imwrite('{}/generated_{}.jpg'.format(save_path, os.path.splitext(img_n)[0]), res)

            except:

                pass

    else:

        output = np.array(autoencoder.model.predict([np.array(img)])[0])

        cv2.imwrite('{}/generated_{}.jpg'.format(save_path, os.path.splitext(img_n)[0]), output)

```