This dataset provides the foundational resources for evaluating and optimizing Formula L , a novel mathematical framework for semantic-driven task allocation in multi-agent systems (MAS) powered by large language models (LLM). The dataset includes Python code and both empirical and synthetic data, specifically designed to validate the effectiveness of Formula L in improving task distribution, contextual relevance, and dynamic adaptation within MAS.
The dataset comprises:
We present the SynSUM benchmark, a synthetic dataset linking unstructured clinical notes to structured background variables. The dataset consists of 10,000 artificial patient records containing tabular variables (like symptoms, diagnoses and underlying conditions) and associated clinical notes describing the fictional patient encounter in the domain of respiratory diseases.
The presented dataset contains information about struts utilized in a material system, including three key attributes: strut diameter, strut type, and sample number. The strut diameter describes the structural element's physical dimension, whereas the strut type specifies the design or placement inside the material, such as edge configurations. A sample number is assigned to each sample, identifying it uniquely. This data can be used in machine learning systems to forecast material qualities, optimize designs, and investigate the effect of strut configurations on structural performance.
ShanghaiTechRGBD is a large-scale RGB-D crowd counting dataset including 2,293 pairs of RGB-D images, which were annotated with the heads of 144,512 pedestrians. In addition, to facilitate the verification of the algorithm, the dataset is divided into training and testing sets, which contain 1,193 pairs of images and 1,000 pairs of images, respectively. This dataset is derived from the following paper: 1. Locating_and_Counting_Heads_in_Crowds_With_a_Depth_Prior. 2. Density map regression guided detection network for rgb-d crowd counting and localization. To use this dataset, please cite:D.
DroneRGBT dataset consists of 3,600 pairs of RGB-T images, which have annotations for a total of 175,698 pedestrians that were captured by drones. This dataset is split into 1,800 pairs of images for training, and the rest are used for testing, thereby further improving the comprehensiveness of our experimental verification. This dataset is derived from the following paper：Rgb-t crowd counting from drone: A benchmark and mmccn network.To use the dataset, please cite: T. Peng, Q. Li, and P.
RGB-T Crowd counting. RGBT-CC dataset is made up of 2,030 pairs of RGB-T images, each having a resolution of 640×480, and these images were captured under different scene circumstances and lighting conditions. A total of 138,389 pedestrians are precisely annotated in this dataset, and there are roughly 68 pedestrians on average in each image. It is divided into three separate subsets: training is performed using 1,030 images, 200 images are used for validation, and 800 images are for testing.
Scene understanding in a contested battlefield is one of the very difficult tasks for detecting and identifying threats. In a complex battlefield, multiple autonomous robots for multi-domain operations are likely to track the activities of the same threat/objects leading to inefficient and redundant tasks. To address this problem, we propose a novel and effective object clustering framework that takes into account the position and depth of objects scattered in the scene. This framework enables the robot to focus solely on the objects of interest.
This dataset contains high-resolution retinal fundus images collected from 495 unique subjects from Eye Care hospital in Aizawl, Mizoram, for diabetic retinopathy (DR) detection and classification. The images were captured over five years using the OCT RS 330 device, which features a 45° field of view (33° for small-pupil imaging), a focal length of 45.7 mm, and a 6.25 mm sensor width. Each image was acquired at a resolution of 3000x3000 pixels, ensuring high diagnostic quality and the visibility of subtle features like microaneurysms, exudates, and hemorrhages.
This dataset was produced as part of the NANCY project (https://nancy-project.eu/), with the aim of using it in the fields of communication and computer vision. Within the dataset are contained three different scenarios, each of which has three videos. All three videos were captured by different devices; a vehicle-mounted unit, a roadside unit (RSU), and a drone.
Sign Language Recognition integrates computer vision and natural language processing to automatically interpret hand gestures and translate them into spoken or written Bengali. The primary goal is to bridge the communication gap between sign language users and non-users by recognizing gestures, movements, postures, and facial expressions that correspond to spoken language elements. Since hand gestures are the cornerstone of sign language communication, they play a pivotal role in improving the accuracy of sign language recognition systems.
