Abstract
With the upgrading of the high-performance image processing platform and visual internet of things sensors, VIOT is widely used in intelligent transportation, autopilot, military reconnaissance, public safety, and other fields. However, the outdoor visual internet of things system is very sensitive to the weather and unbalanced scale of latent object. The performance of supervised learning is often limited by the disturbance of abnormal data. It is difficult to collect all classes from limited historical instances. Therefore, in terms of the anomaly detection images, fast and accurate artificial intelligence-based object detection technology has become a research hot spot in the field of intelligent vision internet of things. To this end, we propose an efficient and accurate deep learning framework for real-time and dense object detection in VIOT named the Edge Attention-wise Convolutional Neural Network (EAWNet) with three main features. First, it can identify remote aerial and daily scenery objects fast and accurately in terms of an unbalanced category. Second, edge prior and rotated anchor are adopted to enhance the efficiency of detection in edge computing internet. Third, our EAWNet network uses an edge sensing object structure, makes full use of an attention mechanism to dynamically screen different kinds of objects, and performs target recognition on multiple scales. The edge recovery effect and target detection performance for long-distance aerial objects were significantly improved. We explore the efficiency of various architectures and fine tune the training process using various backbone and data enhancement strategies to increase the variety of the training data and overcome the size limitation of input images. Extensive experiments and comprehensive evaluation on COCO and large-scale DOTA datasets proved the effectiveness of this framework that achieved the most advanced performance in real-time VIOT object detection.
Highlights
Edge prior reconstruction and attention-wise modules are embedded into the Edge Attention-wise Convolutional Neural Network (EAWNet), which helps in performing efficient latent search and localization (ii) With the combination of intelligent connection and residual link, rotating bounding box, and synthesis loss function, the visual loss of intensive detection is reduced to a minimum
Self-attention modules are adopted to underline the meaningful information of feature maps while disregarding useless information
Comparative experiments are conducted on the benchmark dataset COCO with state-of-theart methods
Summary
Advanced object detection methods have greatly improved over the past few years, and several methods have been introduced to optimize the network structure, which can be divided into single-stage and double-stage; the use of an attention module to improve the efficiency of searching is not well investigated. Single-stage detector: The most representative single-stage detectors are YOLO [5, 6] and SSD [7] They use feature maps to predict the confidence and location of a multitarget receptive field block (RFB) network to achieve accurate and fast object detection. Edge prior reconstruction and attention-wise modules are embedded into the EAWNet, which helps in performing efficient latent search and localization (ii) With the combination of intelligent connection and residual link, rotating bounding box, and synthesis loss function, the visual loss of intensive detection is reduced to a minimum. A rotating bounding box, designed for aerial image object detection, proved to be beneficial for the recognition of dense and tiny objects
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
