Project descriptions
The project focuses on basic image classification methods, utilizing transfer learning techniques by freezing some layers of the EfficientNetB7 model pre-trained on the Imagenet dataset. For the study cases, some layers in the EfficientNetB7 model will be allowed to be retrained for weight updates. The experiment uses a challenging food dataset with 101 food categories and a total of 101,000 food images, including 250 manually reviewed test images and 750 training images for each class. All images were rescaled to a maximum side length of 512 pixels by the dataset's author. You can review the dataset from its official website.