Information-theoretic bounds on target recognition performance based on degraded image data

2002 ◽  
Vol 24 (9) ◽  
pp. 1153-1166 ◽  
Author(s):  
A. Jain ◽  
P. Moulin ◽  
M.I. Miller ◽  
K. Ramchandran
Keyword(s):  
Target Recognition ◽  
Recognition Performance ◽  
Image Data ◽  
Information Theoretic ◽  
Degraded Image

Information-theoretic bounds on target recognition performance

2000 ◽  
Author(s):  
Avinash Jain ◽  
Pierre Moulin ◽  
Michael I. Miller ◽  
Kannan Ramchandran
Keyword(s):  
Target Recognition ◽  
Recognition Performance ◽  
Information Theoretic

scholarly journals Multi-Stream Convolutional Neural Network for SAR Automatic Target Recognition

2018 ◽  
Vol 10 (9) ◽  
pp. 1473 ◽  
Author(s):  
Pengfei Zhao ◽  
Kai Liu ◽  
Hao Zou ◽  
Xiantong Zhen
Keyword(s):  
Feature Fusion ◽  
Target Recognition ◽  
Recognition Performance ◽  
Image Data ◽  
Automatic Target Recognition ◽  
Multiple Views ◽  
Sar Images ◽  
Operating Condition ◽  
High Dependency ◽  
Highly Nonlinear

Despite the fact that automatic target recognition (ATR) in Synthetic aperture radar (SAR) images has been extensively researched due to its practical use in both military and civil applications, it remains an unsolved problem. The major challenges of ATR in SAR stem from severe data scarcity and great variation of SAR images. Recent work started to adopt convolutional neural networks (CNNs), which, however, remain unable to handle the aforementioned challenges due to their high dependency on large quantities of data. In this paper, we propose a novel deep convolutional learning architecture, called Multi-Stream CNN (MS-CNN), for ATR in SAR by leveraging SAR images from multiple views. Specifically, we deploy a multi-input architecture that fuses information from multiple views of the same target in different aspects; therefore, the elaborated multi-view design of MS-CNN enables it to make full use of limited SAR image data to improve recognition performance. We design a Fourier feature fusion framework derived from kernel approximation based on random Fourier features which allows us to unravel the highly nonlinear relationship between images and classes. More importantly, MS-CNN is qualified with the desired characteristic of easy and quick manoeuvrability in real SAR ATR scenarios, because it only needs to acquire real-time GPS information from airborne SAR to calculate aspect differences used for constructing testing samples. The effectiveness and generalization ability of MS-CNN have been demonstrated by extensive experiments under both the Standard Operating Condition (SOC) and Extended Operating Condition (EOC) on the MSTAR dataset. Experimental results have shown that our proposed MS-CNN can achieve high recognition rates and outperform other state-of-the-art ATR methods.

Image stabilization technology for active bridge-vessel anti-collision monitoring system

2021 ◽  
Author(s):  
Zhufeng Shao ◽  
Haiying Ma ◽  
Ye Xia ◽  
Junjie Wang
Keyword(s):  
New Technologies ◽  
Target Recognition ◽  
Image Data ◽  
Tracking Data ◽  
Collision System ◽  
High Quality ◽  
Image Stabilization ◽  
Site Monitoring ◽  
New Research ◽  
Sea Area

<p>In recent years, the active anti-collision system using new technologies such as image target recognition between ship and bridge becomes a new research hotspot. Due to camera jitter, it is not easy to deeply mine the monitoring image data. This paper puts forward an anti-jitter algorithm to obtain the ship monitoring track in the sea area removing the camera jitter. It uses electronic image stabilization, sea-sky line anti jitter filtering, and other methods to process the on-site monitoring video, then compares the effect of each technique, and finally obtains high-quality ship tracking data. Through this method, a high-quality ship monitoring track in the bridge area can be obtained.</p>

scholarly journals Underwater Target Recognition Based on Multi-Decision LOFAR Spectrum Enhancement: A Deep-Learning Approach

2021 ◽  
Vol 13 (10) ◽  
pp. 265
Author(s):  
Jie Chen ◽  
Bing Han ◽  
Xufeng Ma ◽  
Jian Zhang
Keyword(s):  
Feature Extraction ◽  
Deep Learning ◽  
Recognition Accuracy ◽  
Target Recognition ◽  
Signal To Noise Ratio ◽  
Recognition Performance ◽  
Low Frequency ◽  
Learning Approach ◽  
Decision Algorithm ◽  
Underwater Target

Underwater target recognition is an important supporting technology for the development of marine resources, which is mainly limited by the purity of feature extraction and the universality of recognition schemes. The low-frequency analysis and recording (LOFAR) spectrum is one of the key features of the underwater target, which can be used for feature extraction. However, the complex underwater environment noise and the extremely low signal-to-noise ratio of the target signal lead to breakpoints in the LOFAR spectrum, which seriously hinders the underwater target recognition. To overcome this issue and to further improve the recognition performance, we adopted a deep-learning approach for underwater target recognition, and a novel LOFAR spectrum enhancement (LSE)-based underwater target-recognition scheme was proposed, which consists of preprocessing, offline training, and online testing. In preprocessing, we specifically design a LOFAR spectrum enhancement based on multi-step decision algorithm to recover the breakpoints in LOFAR spectrum. In offline training, the enhanced LOFAR spectrum is adopted as the input of convolutional neural network (CNN) and a LOFAR-based CNN (LOFAR-CNN) for online recognition is developed. Taking advantage of the powerful capability of CNN in feature extraction, the recognition accuracy can be further improved by the proposed LOFAR-CNN. Finally, extensive simulation results demonstrate that the LOFAR-CNN network can achieve a recognition accuracy of 95.22%, which outperforms the state-of-the-art methods.

scholarly journals A SAR Target Recognition Method via Combination of Multilevel Deep Features

2021 ◽  
Vol 2021 ◽  
pp. 1-8
Author(s):  
Junhua Wang ◽  
Yuan Jiang
Keyword(s):  
Target Recognition ◽  
Recognition Performance ◽  
Superior Performance ◽  
Recognition Method ◽  
Sar Images ◽  
Noise Interference ◽  
Feature Sets ◽  
Weighted Fusion ◽  
Standard Operating Condition ◽  
Sar Target Recognition

For the problem of synthetic aperture radar (SAR) image target recognition, a method via combination of multilevel deep features is proposed. The residual network (ResNet) is used to learn the multilevel deep features of SAR images. Based on the similarity measure, the multilevel deep features are clustered and several feature sets are obtained. Then, each feature set is characterized and classified by the joint sparse representation (JSR), and the corresponding output result is obtained. Finally, the results of different feature sets are combined using the weighted fusion to obtain the target recognition results. The proposed method in this paper can effectively combine the advantages of ResNet and JSR in feature extraction and classification and improve the overall recognition performance. Experiments and analysis are carried out on the MSTAR dataset with rich samples. The results show that the proposed method can achieve superior performance for 10 types of target samples under the standard operating condition (SOC), noise interference, and occlusion conditions, which verifies its effectiveness.

Image Quality and Target Recognition

1967 ◽  
Vol 9 (1) ◽  
pp. 5-32 ◽  
Author(s):  
Corwin A. Bennett ◽  
Samuel H. Winterstein ◽  
Robert E. Kent
Keyword(s):  
Grain Size ◽  
Image Quality ◽  
Modulation Transfer Function ◽  
Target Recognition ◽  
Recognition Performance ◽  
Aerial Photographs ◽  
Modulation Transfer ◽  
Task Definition ◽  
Subject Selection ◽  
Target Characteristics

The terminology and literature in the area of image quality and target recognition are reviewed. An experiment in which subjects recognized strategic and tactical targets in aerial photographs with controlled image degradations is described. Some findings are: Recognition performance is only moderate for representative conditions. There are wide differences among target types in the recognizability. Knowledge of a target's presence (briefing) greatly aids recognition. Better resolution means better performance. Enlarging the image such that a line of resolution subtends more than three minutes of arc hinders recognition. Grain size should be kept below 20 seconds of arc. It is suggested that the eventual application of the modulation transfer function approach to measurement of image quality and target characteristics will enable a quantitative subsuming of various quality-size relationships. More attention needs to be paid in recognition research to suitable task definition, target description, and subject selection.

Target recognition performance following whole-views, part-views, and both-vlews training

2017 ◽  
pp. 135-141
Author(s):  
Sehchang Hah ◽  
Deborah A. Reisweber ◽  
Jose A. Picart ◽  
Harry Zwick
Keyword(s):  
Target Recognition ◽  
Recognition Performance
Sign in / Sign up

Export Citation Format

Share Document