Anaheim Convention Center
Anaheim, California, United States
26 - 30 April 2020
Conference SI109
Automatic Target Recognition XXX
show | hide
Abstract Due:
16 October 2019

Author Notification:
20 December 2019

Manuscript Due Date:
1 April 2020

show | hide
Conference Chairs
Program Committee
Program Committee continued...
Call for
This conference will emphasize all aspects relating to the modern automatic and machine assisted target and object recognition technology: concepts such as model-based object/target recognition and tracking, neural networks, wavelets, information fusion, knowledge-based methods, adaptive and learning approaches, and advanced signal and image processing concepts for detection, tracking, and recognition for sonar/acoustic, EO, IR, radar, laser radar, multispectral and hyperspectral sensors. Papers dealing with the entire spectrum of algorithms, systems, and architecture in ATR/AOR will be considered.

In particular papers on the model-based solutions will be considered. This includes hypotheses of the initial sets of the sensor data, predictive models of the target features and their relationships, techniques of evaluations/comparisons of the predicted models with the features extracted from the data. Suggested topics also include methods of imputation of missing or sparse data and subsequent evaluation of the results.

Another extremely important challenge for ATR is the evaluation and prediction of ATR performance given the practical limitation that data sets cannot represent the extreme variability of the real world. Methods are sought that allow a rapid insertion of new targets and adaptive algorithms capable of supporting flexible and sustained employment of ATR. A key technical challenge is the development of affordable ATR solutions that employ an open architecture to provide timely hardware and software insertion.

A related topic is the ability to detect and recognize new targets using limited amounts of labeled training imagery. Unlike other applications of computer vision, the ATR problem is characterized by limited amounts of ground truth and labeled training information. Therefore, papers that deal with limited amounts of training data, and the ability to rapidly learn new targets are of significant interest.

Papers are solicited in the following and related topics:

Machine Learning for ATR
  • Deep learning
  • Adversarial learning
  • Multi-view learning
  • Training methodologies.
  • Learning new targets with limited training data
Geospatial Remote Sensing Systems
  • Object recognition from multi-view 3D
  • Object level change detection, recognizing the object from the change
  • Wide area search – finding the object of interest in a scene
  • Scene understanding/Sensemaking – inference of activity from a single image
  • Performance evaluation issues.
IR-based Systems
  • Detection, tracking, and recognition
  • Phenomenological modeling of targets and background
  • Polarization diversity
  • Target/object and scene segmentation
  • Performance evaluation issues.
Hyperspectral-based Systems Registration Issues
  • Detection, tracking, and recognition
  • Phenomenological modeling of targets and background
  • Polarization and waveform adaptation
  • Target/object and scene segmentation
  • Performance evaluation issues.
Radar/Laser Radar-based Systems
  • High-range resolution radar techniques
  • Joint radar target tracking and classification approaches
  • Ultra-wide band radar techniques
  • Doppler, polarization, and waveform diversity for target classification
  • Detection, tracking, recognition, segmentation, target, and clutter modeling
  • Multisensory processing and fusion
  • Performance evaluation issues.
Sonar/Acoustic and Seismic-based Systems
  • Inverse scattering issues
  • Direct scattering of acoustic waves
  • Tomographic image formation
  • Material identification
  • Ultra-wide band methods for target detection and classification
  • Multisensory fusion
  • Biosensor systems
  • Performance evaluation issues.
New Methodologies
  • Information theoretical approaches in ATR
  • Distributed and centralized sensor decision making
  • Model-based object recognition
  • Neural networks for ATR applications
  • Wavelet decomposition methods for ATR
  • Machine learning approaches such as deep learning, transfer learning, dictionary learning and manifold learning applications to ATR
  • Mission adaptive systems
  • Data characterization
  • Performance estimation and modeling
  • ATR/AOR development tools
  • ATR/AOR architecture
  • Algorithms for human detection, tracking, and activity recognition.
Back to Top