Effect of fusing features from multiple DCNN architectures in image classification. Issue 7 (1st July 2018)

• Record Type: Journal Article
• Authors: Akilan, Thangarajah; Wu, Qingming Jonathan; Zhang, Hui
• Is Part Of: IET image processing. Volume 12:Issue 7(2018)
• Page Start: 1102
• Page End: 1110
• Publication Date: 2018-07-01
• Subjects: image classification -- feature extraction -- image representation -- neural nets -- principal component analysis -- image reconstruction -- computer vision; DCNN architectures -- automatic image classification -- feature extraction -- pre‐trained deep convolutional neural networks -- generalised feature space -- principal component reconstruction -- energy‐level normalisation -- fused feature vectors -- image statistics representation -- multiclass linear support vector machine -- FFV -- computer vision; Image processing -- Periodicals; 621.36705
• DOI: 10.1049/iet-ipr.2017.0232
• Languages: English
• ISSNs: 1751-9659

Abstract

Abstract : Automatic image classification has become a necessary task to handle the rapidly growing digital image usage. It has branched out many algorithms and adopted new techniques. Among them, feature fusion‐based image classification methods rely on hand‐crafted features traditionally. However, it has been proven that the bottleneck features extracted through pre‐trained convolutional neural networks (CNNs) can improve the classification accuracy. Thence, this study analyses the effect of fusing such cues from multiple architectures without being tied to any hand‐crafted features. First, the CNN features are extracted from three different pre‐trained models, namely AlexNet, VGG‐16, and Inception‐V3. Then, a generalised feature space is formed by employing principal component reconstruction and energy‐level normalisation, where the features from individual CNN are mapped into a common subspace and embedded using arithmetic rules to construct fused feature vectors (FFVs). This transformation play a vital role in creating a representation that is appearance invariant by capturing complementary information of different high‐level features. Finally, a multi‐class linear support vector machine is trained. The experimental results demonstrate that such multi‐modal CNN feature fusion is well suited for image/object classification tasks, but surprisingly it has not been explored so far by the computer vision research community extensively.