Presentation Topics

Approaches and advances in deep neural networks, including transformer-based networks and state-space models
Vision language models, large multimodal models and large language models — and how to incorporate them into systems and applications
Approaches and advances in object tracking
Conventional (non-ML based) algorithms, including signal/image processing, detection, classification
Event and action recognition (e.g., temporal neural networks)
Improving neural network model accuracy in the real world
Learning at the edge: few-shot learning, continuous learning
Neural network optimization for deployment
Monocular depth estimation
Perception and scene understanding using depth sensors, radar, lidar, etc.
Practical neuromorphic algorithms, including spiking neural networks
Practical vision language models and use cases
Sensor integration and fusion in ML-based systems—vision, depth, audio, haptic, etc.

Practical object tracking approaches
Camera selection and image quality metrics
Ensuring quality data for deep learning: data collection, curation and annotation
Fundamentals of depth sensors: time-of-flight, lidar, stereo, structured light
Fundamentals of optics and image sensors
Fundamentals of training AI models for computer vision applications
Fundamentals of radar for machine perception
Introduction to CNNs
Introduction to ML for visual perception
Introduction to and survey of DNN-based object detectors and use cases
Introduction to segmentation: semantic, instance, panoptic
Introduction to transformers
Introduction to large language models
Introduction to vision language models (and potentially other large multimodal models)
Introduction to state-space sequence modeling
Introduction to model compression approaches—quantization, pruning, sparsity, distillation, etc.
Model training problems, fixes, insights, best practices
Introduction to SLAM

Case studies of end-to-end computer vision and AI-based solutions, pitfalls and insights
Cloud-edge tradeoffs
End-to-end system optimization
Event-based and neuromorphic sensors and processors
Identifying and mitigating bias in AI
Image and non-image sensors and processing techniques: lidar, radar, time-of-flight, stereo, audio
Insights and best practices in selecting AI/CV technologies and techniques for real-world products
Insights and best practices in tuning and adapting vision language models
Low-power, low-cost computer vision and AI components—SoCs, NPUs, etc.
Real-world performance metrics for developing and maintaining AI-based products
Training data collection, curation, annotation, generation, augmentation
MLOps—training a model, best practices, insights
MLOps—deploying and maintaining models in production
MLOps—testing a model in the lab vs. in the wild, best practices, insights and explainability
Image sensors, optics, camera selection, image quality metrics
Standards
Tools, frameworks and techniques for developing and deploying AI-based products
Top-to-bottom software stacks for AI-based systems and applications
Building applications with vision language models, large multimodal models and large language models
Insights and best practices for fine-tuning task-specific LMMs
Using AI agents in products

Call for Proposals Suggested Topics