Title of the special session: - Advancing Border Security with AI, Laser-Based Detection, and Enhanced Monitoring Solutions
Proposed by: - Border Security Force (BSF), Jalandhar, Punjab
Themes:
1. Countering Drone-Based Smuggling and Security Threats
Description: Drones are increasingly used by anti-national elements to smuggle drugs (e.g., heroin) and weapons (e.g., pistols, IEDs) across the border. To combat this, an advanced anti-drone system is needed that:
-Uses AI-powered detection, identification and tracking to neutralise drones in real time.
-Integrates radar, infrared, and RF sensors for all-weather drone detection and tracking.
-Deploys countermeasures such as jamming, laser-based neutralization, GPS spoofing or AI-assisted interception.
2. AI-Enabled Processing of Surveillance Data for Automated Threat Detection
Description: The Border Security Force (BSF) uses an extensive network of surveillance cameras. However, the vast amount of video data requires real-time AI processing to reduce human dependency and improve response time. There is a need for an AI-powered surveillance system working as an additional layer on EO Cameras that:
-Automatically detects and classifies threats such as unauthorized human movement, smuggling activities, and drone intrusions.
-Filters out non-threat activities, reducing false alarms and ensuring efficient resource allocation.
-Provides real-time alerts with geo-tagging, improving situational awareness and response time.
3. Laser-Based Intrusion Detection for Enhanced Border Security
Description: Traditional fencing and visual monitoring have limitations, especially in low-light and foggy conditions. A laser-based intrusion detection system is required that:
-Creates an invisible laser wall across vulnerable points of the border.
-Instantly detects breaches by analysing disruptions in the laser path.
-Integrates with AI-based analytics to differentiate between human intrusion and non-threatening disturbances like animals.
4. Fog-Resistant Surveillance Technology
Description: Dense fog during winters reduces visibility, making border security vulnerable. A fog-resistant surveillance system should:
-Utilize multi-spectral imaging (thermal, infrared, or radar-based AI) to detect movement.
-Enhance real-time object classification to distinguish between threats and non-threats.
-Operate autonomously and provide instant alerts to quick reaction teams and troops for rapid deployment.
5. AI-Driven Database Management System for Smugglers and Anti-National Elements:
Description: Currently, information about captured and suspected smugglers/terrorists is scattered, making it difficult to track patterns and repeat offenders. There is a need for an intelligent database management system that:
-Consolidates and updates records of smugglers and anti-national elements operating along the border augmented with face recognition capabilities.
-Uses AI to analyse behavioural patterns, identifying networks and potential threats.
-Provides real-time access to field officers for quick decision-making and intelligence sharing.
Expected Outcomes:
1. Anti-drone system to combat drone menace on the border.
Title of the special session: - Deep Learning for Limited-label Data in Critical Applications Names of proposers: - Dr Jignesh S. Bhatt*, Prof Dharmendra Singh* Affiliations: - *Indian Institute of Information Technology Vadodara (Gujarat) Email addresses: - jignesh.bhatt@iiitvadodara.ac.in*; dharmfec@iiitvadodara.ac.in** Description of the special session/workshop: -
Deep learning has been the driving technology behind many data-driven technological developments in the last two decades. We have witnessed state-of-the-art algorithms with availability of large labeled data and increased compute power. Nevertheless, many critical areas are still found limited with availability of data while processing is resource constrained. It mainly includes remote sensing, defense, healthcare, and agricultural applications. Hence, it has given rise to newer deep learning paradigms and architectures for limited-label data learning.
In this special session, we focus on such cutting edge deep learning technological developments targeting critical applications in sensor data generated from aircraft or spacecraft, drones, use of life critical sensors in medical imaging, and agricultural applications making them AI-enabled. The session would begin with a refresher in machine learning and shall embark on the high-end technical details for self-supervised learning, reinforcement learning, generative AI, as well as development of analytical deep models. Our focus shall be on large scale image/video datasets captured using different sensors in the said application domains. We present both theoretical and applied approaches to benefit a larger audience including students, researchers, and practitioners. Topics of Interest:
A broad outline of the session would be as follows: ● The problem of learning from data
○ Evolution to current state
● The problem of learning from limited-label data
○ Evolution to current state
● Case studies
○ Hyperspectral Unmixing
○ Drone-based AI for Agri Apps
○ Multispectral Registration and Fusion
○ Automatic Target Detection in SAR
○ Lung Health Visualization System
○ GenAI for Healthcare
○ Federated Medical Model ● Theoretical Aspects in Deep Networks
● Conclusions and Lessons Learnt
Title of the special session: - Human Machine Teaming: AI Integrated Electro-Optical Systems for Surveillance Names of proposers: - Dr Mr JP Singh, Associate Director (AI), Scientist G and Dr Vaibhav Gupta, Ph.D (Neural Networks), Sr Technical Officer Affiliations: - *IRDE, DRDO, Dehradun Email addresses: - vaibhav.drdo@gmail.com Description of the special session/workshop: -
Topics of Interest: ● AI Integrated Tracking & Detection System for surveillance ● Data Collection, Annotation, Augmentation Techniques in AI Modeling
The Microsoft CMT service was used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support.
