Dementia affects the patient’s memory and leads to language impairment. Research has demonstrated that speech and language deterioration is often a clear indication of dementia and plays a crucial role in the recognition process. Even though earlier studies have used speech features to recognize subjects suffering from dementia, they are often used along with other linguistic features obtained from transcriptions. This study explores significant standalone speech features to recognize dementia. The primary contribution of this work is to identify a compact set of speech features that aid in the dementia recognition process. The secondary contribution is to leverage machine learning (ML) and deep learning (DL) models for the recognition task. Speech samples from the Pitt corpus in Dementia Bank are utilized for the present study. The critical speech feature set of prosodic, voice quality and cepstral features has been proposed for the task. The experimental results demonstrate the superiority of machine learning (87.6 percent) over deep learning (85 percent) models for recognizing Dementia using the compact speech feature combination, along with lower time and memory consumption. The results obtained using the proposed approach are promising compared with the existing works on dementia recognition using speech.

Decoding the intentions of passengers and other road users remains a critical challenge for autonomous vehicles and intelligent transportation systems. Hand gestures are key in these interactions, offering a direct communication channel. Moreover, egocentric videos mimic a first-person perspective, aligning closely with human visual perception. Yet, the development of deep learning algorithms for detecting egocentric hand gestures in autonomous driving is hindered by the absence of useful datasets. Furthermore, there is a pressing need for gesture recognition methods to evolve from CNN-based architectures to transformer models. To address these challenges, we present EgoDriving, a novel dataset of egocentric hand gestures, curated for driving-related hand gestures. Finally, we introduce EgoFormer, an efficient video transformer for egocentric hand gesture classification that is optimized for edge computing deployments. EgoFormer incorporates a Video Dynamic Position Bias (VDPB) module to enhance long-range positional awareness and leverage absolute positions from convolutional sub-layers within its transformer blocks. Designed for low-resource settings, EgoFormer offers substantial reductions in inference latency and GPU utilization while maintaining competitive accuracy against state-of-the-art hand gesture recognition frameworks.