Fdct Frequencyaware Decomposition And Crossmodal Tokenalignment For Multisensor Target Classification

• doi: 10.1109/TAES.2025.3550474
• title: FDCT: Frequency-Aware Decomposition and Cross-Modal Token-Alignment for Multi-Sensor Target Classification
• publisher: IEEE
• isbn:
• issn: 2371-9877
• rank: 2456
• access_type: LOCKED
• content_type: Early Access Articles
• abstract: In automatic target recognition (ATR) systems, sensors may fail to capture discriminative, fine-grained detail features due to environmental conditions, noise created by CMOS chips, occlusion, parallaxes, and sensor misalignment. Therefore, multi-sensor image fusion is an effective choice to overcome these constraints. However, multi-modal image sensors are heterogeneous, and have domain and granularity gap. In addition, the multisensor images can be misaligned due to intricate background clutters, fluctuating illumination conditions, and uncontrolled sensor settings. In this paper, to overcome these issues, we decompose, align, and fuse multiple image sensor data for target classification. We extract the domain-specific and domain-invariant features from each sensor data. We propose to develop a shared unified discrete token (UDT) space between sensors to reduce the domain and granularity gap. Additionally, we develop an alignment module to overcome the misalignment between multi-sensors and emphasize the discriminative representation of the UDT space. In the alignment module, we introduce sparsity constraints to provide a better cross-modal representation of the UDT space and robustness against various sensor settings. We achieve superior classification performance compared to single-modality classifiers and several state-of- the-art multi-modal fusion algorithms on four multi-sensor ATR datasets. Furthermore, the experimental results demonstrate that our multi-sensor classifier outperforms state-of-the-art single modality classifiers, e.g., EfficentNet, ViT, and ConvNeXt, in the VEDAI ATR dataset. Our code will be released at https://github.com/shoaib-sami/FDCT-Multi-Sensor-Classifier/.
• article_number: 10924416
• pdf_url: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10924416
• html_url: https://ieeexplore.ieee.org/document/10924416/
• abstract_url: https://ieeexplore.ieee.org/document/10924416/
• publication_title: IEEE Transactions on Aerospace and Electronic Systems
• conference_location:
• conference_dates:
• publication_number: 7
• is_number: 7778228
• publication_year: 2025
• publication_date:
• start_page: 1
• end_page: 23
• citing_paper_count: 0
• citing_patent_count: 0
• download_count: 31
• insert_date: 20250312
• index_terms:
  • ieee_terms:
    • Sensors
    • Feature extraction
    • Sensor fusion
    • Radar
    • Classification algorithms
    • Convolutional neural networks
    • Sensor phenomena and characterization
    • Object detection
    • Laser radar
    • Transformers
  • author_terms:
    • Automatic Target Recognition
    • Multi-Sensor Fusion
    • Invertible Neural Network
    • Cross-Modal Token Alignm
  • dynamic_index_terms:
    • Target Class
    • Superior Performance
    • Detailed Characterization
    • Detailed Features
    • Classification Performance
    • Sensor Data
    • Single Class
    • Image Sensor
    • Camera Sensor
    • Target Recognition
    • Multiple Sensors
    • Fusion Algorithm
    • Multi-sensor Fusion
    • Domain-invariant Features
    • Alignment Module
    • Domain-specific Features
    • Domain-specific Characteristics
    • Multimodal Sensor
    • Automatic Target Recognition
    • Convolutional Neural Network
    • Object Detection
    • Cross-entropy Loss
    • Image Pairs
    • LiDAR Sensor
    • Low-frequency Features
    • Low-frequency Characteristics
    • Vision Sensors
    • Latent Space
    • Embedding Space
    • Vision Transformer
    • Unmanned Aerial Vehicles
    • Ultrasonic Sensors
    • Ultrasonic Detection
    • Alignment Strategy
    • Alignment Scheme
    • Tractor
    • Radar Sensor
• authors:
  • Author Name: SHOAIB MERAJ SAMI
    Affiliation: LCSEE Dept., West Virginia University, WV, USA
    Author URL: https://ieeexplore.ieee.org/author/37088831183
    ID: 37088831183
    Order: 1
    Author Affiliations:
    • LCSEE Dept., West Virginia University, WV, USA
  • Author Name: MD MAHEDI HASAN
    Affiliation: LCSEE Dept., West Virginia University, WV, USA
    Author URL: https://ieeexplore.ieee.org/author/37088393653
    ID: 37088393653
    Order: 2
    Author Affiliations:
    • LCSEE Dept., West Virginia University, WV, USA
  • Author Name: NASSER M. NASRABADI
    Affiliation: LCSEE Dept., West Virginia University, WV, USA
    Author URL: https://ieeexplore.ieee.org/author/37087365978
    ID: 37087365978
    Order: 3
    Author Affiliations:
    • LCSEE Dept., West Virginia University, WV, USA
  • Author Name: RAGHUVEER RAO
    Affiliation: Army Research Laboratory, MD, USA
    Author URL: https://ieeexplore.ieee.org/author/37281258600
    ID: 37281258600
    Order: 4
    Author Affiliations:
    • Army Research Laboratory, MD, USA

Image Sensor

• sensor_type: CMOS
• resolution: Not stated
• dynamic_range: Not stated
• pixel_size: Not stated
• dark_current: Not stated

Optical Data

• focal_length: Not stated
• aperture: Not stated
• field_of_view: Not stated
• distortion: Not stated

Performance Metrics

• frame_rate: Not stated
• signal_to_noise_ratio: Not stated
• sensitivity: Not stated
• shutter_speed: Not stated
• power_consumption: Not stated
• noise: Not stated

Applications & Benefits

• cell_imaging: Not stated
• benefits: CMOS image sensors enable digital imaging in a variety of applications such as smartphones, medical devices, and automotive systems.

Supporting Organizations

• supported_by: Not stated

Manuscript Details

• publication_date: 2025-03-11

Relevancy Score

• score: 10
• missing_fields:
  • resolution
  • dynamic_range
  • pixel_size
  • dark_current
  • focal_length
  • aperture
  • field_of_view
  • distortion
  • frame_rate
  • signal_to_noise_ratio
  • sensitivity
  • shutter_speed
  • noise
  • cell_imaging
  • supported_by

Processed JSON Filename

request_23f07a38-3893-4466-990b-d136b6e32d53-fdct_frequencyaware_decomposition_and_crossmodal_tokenalignment_for_multisensor_target_classification.json