VIP Cup 2025 at ICIP 2025

IEEE ICIP 2025 Website | 14-18 September 2025 | 2025 VIP Cup Official Document
[Sponsored by the IEEE Signal Processing Society]
 

Unmanned Aerial Vehicles (UAVs), commonly referred to as drones, have gained significant prominence in recent years due to their diverse applications in areas such as surveillance, delivery, agriculture, logistics, disaster response, and military operations. However, the proliferation of drones has also introduced critical challenges such as unauthorized aerial activity, potential threats from payload delivery systems including payload delivery for malicious purposes as well as major security and privacy concerns. This necessitates the detection and tracking of drones, along with the identification of their payloads to ensure safety and security, particularly in sensitive or restricted areas. 

Traditional vision-based approaches for drone detection predominantly rely on RGB images, which are often hindered by environmental factors such as low light, fog, or glare. On the other hand, infrared (IR) imaging provides complementary information by capturing thermal signatures enabling robust detection in challenging conditions such as nighttime or occlusion even in cases where the drones are exceedingly far from the field of view (FoV) of the surveillance cameras. Although it may seem that infrared imaging outperforms RGB imaging in such cases, standalone use of IR data may lack the spatial and textural richness provided by RGB images. This makes it essential to combine the two modalities to leverage their complementary strengths. 

Alongside drone detection, the ability to detect and identify payloads carried by drones is critical due to the potential risks associated with unauthorized or malicious payload delivery. Payloads may include hazardous materials, surveillance equipment, or contraband, posing threats to public safety, security, and privacy. Identifying payloads in real-time enables proactive measures to mitigate risks making it an essential aspect of drone monitoring systems. However, identifying and analyzing payloads carried by drones poses unique challenges. Payloads may vary in size, shape and thermal properties making it difficult to rely on a single modality for robust detection. RGB imaging provides critical visual cues for recognizing payload shapes and textures while IR imaging can highlight heat signatures of payloads, particularly those carrying heat-generating devices or components. A fusion-based approach is therefore crucial to enhance payload detection accuracy and reliability.

Task Description

The datasets, for both payload identification and drone detection, would be made available to the registered participants. The proposed solutions must be able to detect and track drones and identify the payload in real-time for each scenario considered in the datasets. A detailed description of the datasets along with a summary of each scenario considered will be provided with the datasets. Annotations for the training set will be made available along with the release of the datasets. The participants are also required to submit readable code of the proposed algorithm (preferably Python) with appropriate documentation and a brief description of the steps followed. Each solution should contain a demo code that could be used to run the submitted solution for a test video. The classification accuracy for the drone must be displayed in real-time on the test video while it is being played. The classification accuracy for the identified payload can be reported at the end of inference. The participants are also required to share the inference time (an evaluation metric) of their code and system specifications on which it was implemented.

The proposed solution must meet the following criteria:

• The provided dataset of RGB and IR images must be used to train 3 different models - One on each RGB & IR, and one on the dataset obtained from fusion of these image pairs
• The following rules apply for each of the above-mentioned models:
• The model must be able to identify drones and differentiate them from other flying entities in real-time under different scenarios considered in the dataset.
• The proposed solution should be able to track the trajectory of drones under various topographical conditions including distortions considered in the dataset and must be able to make conclusions on whether the drones are approaching or receding from the FoV of the camera source.
• The model must be able to identify the type of payload along with the drone and report the same during the inference. A brief description of different payloads will be made available along with the release of the dataset.

Full technical details, dataset(s), evaluation metrics, and all other pertinent information about the competition is located in the 2025 VIP Cup Official Document.

Important Dates

• Challenge announcement 30 March 2025
• Availability of Training Data – 10th April 2025
• Team Registration Deadline: 30 April 2025 (Register here)
• Availability of Test Data – 20th May 2025
• Final submission due: 12 July 2025
• Finalists announcement: 15 July 2025
• Presentation by top 3 teams and announcement of winners at ICIP 2025: September 14-17, 2025

Registration and Important Resources

Official VIP Cup Team Registration

• All teams MUST be registered through the official competition registration system before the deadline in order to be considered as a participating team. Teams also MUST acknowledge, agree to the SPS Student Terms and Conditions, and meet all eligibility requirements at the time of team registration as well as throughout the competition. The Agreement Form can be found in the Official Terms & Conditions document linked in the top section of this page.
• Registration Link: Register your team for the 2025 VIP Cup before 30 April 2025 and submit work before 12 July 2025!
• 2025 VIP Cup Information page (This is not the official registration, you must still register through the official team registration system to be eligible for prizes and participate in the competition: https://www2.securecms.com/VIPCup/VIPRegistration.asp)
   

Competition Organizers

SPS Technical Committee: Image, Video, and Multidimensional Signal Processing

• Vishnu Ravishankar, Software Development Engineer, ARTPARK, Indian Institute of Science, India
• Dharini Raghavan, Graduate Student, Georgia Institute of Technology, Atlanta, USA
• Dr S Sethu Selvi, Professor, Department of ECE, Ramaiah Institute of Technology, India
• Dr Raghuram S, Associate Professor, Department of ECE, Ramaiah Institute of Technology, India
• Dr. Sitaram Ramachandrula, Senior Director, Data Science, [24]7.ai, India
• Shefali Singh, Student Department of ECE, Ramaiah Institute of Technology, India
• Sangeetha Kar, Student Department of ECE, Ramaiah Institute of Technology, India
• Suman Jangid, Student Department of ECE, Ramaiah Institute of Technology, India

Finalist Teams

Contacts

Competition Organizers (technical, competition-specific inquiries): 

• Vishnu Ravishankar, Dharini Raghavan, Dr S Sethu Selvi, Dr Raghuram S, Dr. Sitaram Ramachandrula, Shefali Singh, Sangeetha Kar, Suman Jangid
• SPS Staff (Terms & Conditions, Travel Grants, Prizes): Jaqueline Rash, SPS Membership Program and Events Manager
• SPS Student Services Committee: Lucas Thomaz, Chair