Track 1: Edge AI for Imaging & Embedded Vision

· On-device visual inference (classification, detection, segmentation)

· Real-time video analytics under latency constraints

· Model compression: pruning, quantization, distillation

· Efficient transformers and lightweight vision backbones

· Edge tracking, re-identification, and multi-camera analytics

· Event cameras and neuromorphic vision at the edge

· Resource-aware scheduling for vision workloads

· Benchmarking edge vision (accuracy–latency–power tradeoffs)

 

 

 

 

Track 2: Computational Imaging & Imaging Pipelines at the Edge

· Edge-based image reconstruction (CT/MRI/ultrasound/optical)

· Denoising, super-resolution, and enhancement on-device

· Learned imaging pipelines (ISP, RAW-to-RGB, HDR)

· Compressive sensing and low-light imaging methods

· Imaging system calibration, correction, and robustness

· Edge acceleration for reconstruction (GPU/NPU/FPGA)

· Adaptive sensing and closed-loop imaging at the edge

· Quality metrics and validation for computational imaging

 

 

 

 

Track 3: Multimodal Sensing & Sensor Fusion for Edge Perception

· Camera–LiDAR–radar fusion architectures

· Time synchronization and calibration for multimodal systems

· 3D perception and mapping on edge platforms

· Edge SLAM and localization for mobile systems

· Audio-visual and tactile/IMU integration at the edge

· Uncertainty-aware fusion and reliability estimation

· Edge analytics for industrial sensing networks

· Dataset creation and evaluation for multimodal perception

 

 

 

 

Track 4: Remote Sensing, UAV, and Geospatial Edge Analytics

· Edge AI for UAV/drone imaging and surveillance

· Satellite/hyperspectral processing near-sensor

· On-board change detection and rapid scene understanding

· Edge geospatial segmentation and object extraction

· Efficient transmission and compression for remote sensing streams

· Edge fusion of EO/IR/SAR and multi-source geospatial data

· Field deployment case studies (agriculture, environment, disaster)

· Trustworthy geospatial analytics (bias, drift, reliability)

 

 

 

 

Track 5: Intelligent Edge Services & Cloud–Edge Orchestration for Sensing Workloads

· Service-oriented architectures for imaging/sensing pipelines

· Stream processing and event-driven edge services

· Cloud–edge collaboration for perception and analytics

· Federated / split learning for sensor and imaging networks

· Resource allocation, autoscaling, and workload placement

· Edge MLOps for sensing systems (deployment, monitoring, updates)

· Digital twins for sensor systems and edge services

· Industrial platforms and real-world deployments (smart factories/cities)

 

 

 

 

Track 6: Trustworthy, Secure, and Efficient Edge Intelligence for Imaging/Sensing

· Privacy-preserving analytics (on-device, federated, anonymization)

· Security for edge AI pipelines (attacks, defenses, secure updates)

· Robustness to noise, occlusion, domain shift, and drift

· Explainable AI for imaging and sensing decisions

· Safety assurance for perception in critical systems

· Energy-efficient computing and thermal-aware operation

· Reliability engineering and fault tolerance for edge devices

· Responsible AI and governance for sensing applications