Fine-grained Image Classification Using Combined DCNN and SVM

Abstract

Fine-grained image classification is usually concerned with the same category of objects, for example different kinds of dogs such as Labrador and Golden Retriever. Since the fine-grained categories belong to the same main category, the gap between the categories is very small. These subtle gaps are easily covered by factors such as color, light and background, which leads to the difficulty of fine-grained image classification. In this paper, we propose a dogs breed recognizer that combined deep convolutional neural network with linear support vector machine to tackle fine-grained image classification problems. Deep convolutional neural network is used as a fixed feature extractor, and linear support vector machine is used as a classifier. The proposed method is tested on a dog dataset including 133 dog breeds and 8,351 images. Experimental result demonstrates that the classification accuracy of proposed method outperforms other conventional features methods by over 10 percentage point.