Selective Feature Connection Mechanism: Concatenating Multi-layer CNN Features with a Feature Selector

Abstract

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.

The paper is under consideration at Pattern Recognition Letters

References in corpus (19)

• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• Striving for Simplicity: The All Convolutional Net
• Wide Residual Networks
• ParseNet: Looking Wider to See Better
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• DRAW: A Recurrent Neural Network For Image Generation
• FractalNet: Ultra-Deep Neural Networks without Residuals
• Highway Networks
• Deformable Convolutional Networks
• Efficient Architecture Search by Network Transformation
• Deep Networks with Stochastic Depth
• COCO-Text: Dataset and Benchmark for Text Detection and Recognition in Natural Images
• Scene Text Detection via Holistic, Multi-Channel Prediction
• HyperNet: Towards Accurate Region Proposal Generation and Joint Object Detection
• EAST: An Efficient and Accurate Scene Text Detector
• Deeply-Fused Nets
• Single Shot Text Detector with Regional Attention
• WordSup: Exploiting Word Annotations for Character based Text Detection
• SRM : A Style-based Recalibration Module for Convolutional Neural Networks

Cited by in corpus (1)

• Industrial object, machine part and defect recognition towards fully automated industrial monitoring employing deep learning. The case of multilevel VGG19