Moid Sandhu | Searcharxiv

arXiv2026

An IMU Dataset for Human Activity Recognition to Support Independent Living in Smart Homes (IMU-HAR-IL)

This document introduces HAR-IMU-IL, a dataset developed for human activity recognition (HAR) using inertial measurement unit (IMU) sensors within a smart home environment with a focus to support objective functional assessment of older adults' independent living (IL). In particular, HAR-IMU-IL includes recordings of 50 participants performing 17 clinically relevant activities of daily living, spanning 4 functional domains essential for independent living: mobility, hygiene, nutrition and hydration, and medication intake. The dataset was collected using 30 IMU sensors, comprising both wearable and object-mounted devices integrated within a real-world residential setting. The dataset includes multi-sensor inertial data captured under realistic, unconstrained conditions, together with detailed annotations ensuring high temporal accuracy and consistency across sensors. A comprehensive data collection protocol was implemented to preserve ecological validity and enable reliable multi-sensor synchronisation. HAR-IMU-IL provides a large-scale, functionally grounded resource for advancing and benchmarking machine learning and artificial intelligence approaches for HAR in home settings. We further demonstrate its utility by developing models capable of accurately recognising both activities and broader functional domains using wearable and object-mounted sensors. These capabilities highlight the dataset's potential to enable applications in continuous activity monitoring, functional health assessment, smart home automation, and assistive technologies that support independent living.

cs.HC↗

arXiv2026

Adaptive Hybrid Optimizer based Framework for Lumpy Skin Disease Identification

Lumpy Skin Disease (LSD) is a contagious viral infection that significantly deteriorates livestock health, thereby posing a serious threat to the global economy and food security. Owing to its rapid spread characteristics, early and precise identification is crucial to prevent outbreaks and ensure timely intervention. In this paper, we propose a hybrid deep learning-based approach called LUMPNet for the early detection of LSD. LUMPNet utilizes image data to detect and classify skin nodules -- the primary indicator of LSD. To this end, LUMPNet uses YOLOv11, EfficientNet-based CNN classifier with compound scaling, and a novel adaptive hybrid optimizer. More precisely, LUMPNet detects and localizes LSD skin nodules and lesions on cattle images. It exploits EfficientNet to classify the localized cattle images into LSD-affected or healthy categories. To stabilize and accelerate the training of YOLOv11 and EfficientNet hybrid model, a novel adaptive hybrid optimizer is proposed and utilized. We evaluate LUMPNet at various stages of LSD using a publicly available dataset. Results indicate that the proposed scheme achieves 99% LSD detection training accuracy, and outperforms existing schemes. The model also achieves validation accuracy of 98%. Moreover, for further evaluation, we conduct a case study using an optimized EfficientNet-B0 model trained with the AdamW optimizer, and compare its performance with LUMPNet. The results show that LUMPNet achieves superior performance.

cs.CV↗