Selective feature connection mechanism: Concatenating multi-layer CNN features with a feature selector

作     者：Du, Chen Wang, Chunheng Wang, Yanna Shi, Cunzhao Xiao, Baihua 

作者机构：Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China Univ Chinese Acad Sci Beijing Peoples R China 

出 版 物：《PATTERN RECOGNITION LETTERS》 (模式识别快报)

年 卷 期：2020年第129卷

页      面：108-114页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Key Programs of the Chinese Academy of Sciences [ZDBS-SSWJSC003, ZDBS-SSW-JSC004, ZDBS-SSWJSC005] National Natural Science Foundation of China (NSFC) [61601462, 61531019, 71621002] 

主　　题：Feature combination Network architecture Selective feature connection mechanism Convolutional neural network 

摘      要：Different layers of deep convolutional neural networks(CNNs) can encode different-level information. High-layer features always contain more semantic information, and low-layer features contain more detail information. However, low-layer features suffer from the background clutter and semantic ambiguity. During visual recognition, the feature combination of the low-layer and high-level features plays an important role in context modulation. If directly combining the high-layer and low-layer features, the background clutter and semantic ambiguity may be caused due to the introduction of detailed information. In this paper, we propose a general network architecture to concatenate CNN features of different layers in a simple and effective way, called Selective Feature Connection Mechanism (SFCM). Low-level features are selectively linked to high-level features with a feature selector which is generated by high-level features. The proposed connection mechanism can effectively overcome the above-mentioned drawbacks. We demonstrate the effectiveness, superiority, and universal applicability of this method on multiple challenging computer vision tasks, including image classification, scene text detection, and image-to-image translation. (C) 2019 Elsevier B.V. All rights reserved.