State Classification of Cooking Objects using pre-trained VGG19 Network
Anupriya Kaushik · 2019
Object and state recognition are both parts of the image classification problem, in which the object recognition is the process of identifying an object in an image whereas the state recognition is the process of identifying the progress by recognizing the state of object in that particular image.Object and state recognition inside the kitchen is very important if the cooking is being carried out by a robot.In this article the approach which is used to classify the object among one of the total 11 states is transfer learning.A dataset with a large number of images was maintained which consisted of various objects which were divided into 11 different states.A new network has been created by adding few more layers to the pre trained network called VGG-19 which is used as the base model.This paper discusses about how to improve the accuracy for the state recognition by using transfer learning.