No abstract

No abstract

It is an ingrained ability of humans to recognize and classify an image within a millisecond. This is because since our childhood, the human brain is accustomed to seeing a variety of images from the same category. However image classification in computers is a challenging process. To train computers to recognize and categorize images to a specific category, thousands of images of the same category must be sent, by which the computers can figure out and store the pattern from all the images of that specific category. When an image of the same category is sent again, it will easily recognize the image belongs to a specific class based on the patterns that are stored for that class. The objective of this paper is to explore the different transfer learning techniques that can be used for image classification task with high accuracy. Keywords: VGG19, ResNet, Densenet, Inceptionv3, Xception