Smart City Surveillance Unveiling Indian Person Attributes in Real Time (SCSPAR 2.0)

Organisers: Dr. Renu M. Rameshan, Dr. Shikha Gupta, Anurag Bajpai - (Vehant Technologies)

Person attribute recognition (PAR) is crucial for city surveillance as it enables the identification and tracking of individuals across multiple cameras. It also gives the system the ability to retrieve instances that have specific attributes, a crucial requirement in surveillance applications. Typical attributes include gender, clothing style, carrying items, etc, which provide high-level semantic information. Existing standard datasets like PETA, Market 1501, and PA100K lack attribute classes relevant to Indian attire, such as kurta, salwar, dupatta, and saree, etc, which are essential for the Indian context. Several solutions are also available in the literature which gives good results on these datasets. But these solutions are not directly applicable to the Indian scenario where there is a change in skin color, dressing style, etc. Even the class information needs to be modified to suit the Indian scenario. The proposed challenge addresses this gap by focusing on detecting person attributes explicitly tailored for the Indian scenario, enhancing the accuracy and relevance of attribute recognition in smart city environments. We will be providing a sample dataset with the intent of sensitizing the participants of the Indian scenario. Participants are encouraged to enhance the dataset to meet their training requirements in a suitable manner.

RACD Challenge 2025: Human Segmentation, Background Removal, and Super-Resolution Enhancement in Residential Surveillance Flyer

Organisers: Dr. Bindu V R (Mahatma Gandhi University), Dr. Deepak Mishra (Indian Institute of Space Science and Technology), Dr. Abdul Jabbar P, Ms. Nisha Shamsudin (Mahatma Gandhi University), Ms. Mintu Movi (Mahatma Gandhi University)

Residential surveillance has become a vital aspect of modern smart home systems, providing essential security and behavioral insights through continuous monitoring. However, the effectiveness of such surveillance largely depends on the ability to accurately detect and analyze human presence, even under challenging conditions such as low-resolution footage, varying lighting, occlusions, and complex background environments. The Residential Activity Capture Dataset (RACD) was developed as a benchmark resource to address these challenges and support research in human segmentation, background removal, and human- focused super-resolution.

RASOI – Recognition and Segmentation of Indian Thali using AI

Organisers: Ashoka University

Food recognition plays a key role in diet tracking, calorie estimation, healthcare, and personalized nutrition. Though there exists a lot of work on other cuisines; Indian food, especially the composite Thali is underrepresented. RASOI (Recognition and Segmentation of Indian Thali), as a challenge aims to push the frontiers of AI by identifying each food item in an Indian Thali, which is visually complex and culturally diverse.

DeHaDo-AI: Challenge on Deciphering Handwritten Documents using Robust AI Models

Organisers: Dr. Sasithradevi A (VIT Chennai), Dr. Prakash P, Professor (MIT Campus, Chennai), Dr. S Md Mansoor Roomi (TCE Madurai), Mr. D. Sabarinathan (Couger Inc., Japan).

The DeHaDo-AI Challenge introduces a dataset of scanned handwritten application forms that reflect real-world variability in handwriting styles, ink quality, image clarity, and form structure. The dataset specifically focuses on English handwritten documents authored by Indian citizens, capturing diverse regional writing styles and document formats. This challenge invites participants to develop robust AI models capable of accurately recognizing handwritten English text, validating field-level entries, and verifying form completeness. The solutions developed through our challenge can be applied to automate data entry in government services, banking, education, and healthcare sectors. They also support digital archiving and retrieval of legacy handwritten records.