Search

Human Activity Recognition (HAR) using smartphone sensors has been identified as a significant emerging research domain. Its application areas exhibit the performance from the intelligent tailored activity monitoring. Researchers have proposed various HAR models to recognize the human activity patterns using traditional smartphone sensor data. In addition, embedding contextual information such as data availability, sensing device orientation, body part location, axis layout, and more makes a fruitful impact on the quality of activity sensor data. The research challenge occurred due to the lack of contextual information in sensor data, leading to activity patterns ambiguity. Often, the motion sensor was separately used to acquire contextual information that consumes unnecessary computational resources. In this paper, we have used activity sensor data availability as contextual information. The proposed Deep Context Model (DCM) recognizes the activity pattern in the dual context-attention mode, i.e., static and dynamic context. The proposed model consists of convolutional and recurrent networks that find the associated activity patterns in a dual context. The convolutional networks are excellent for automatic feature extraction in a static context, whereas recurrent networks are used in a dynamic context for memorizing patterns. We evaluated the performance of the proposed model using the open accessible KU-HAR dataset. The experimental outcomes revealed that DCM has achieved the F1 score of 98.96% and 99.62% for static and dynamic context, respectively. Further, the robustness and applicability of the proposed model have been gauged using the HHAR dataset. [ABSTRACT FROM AUTHOR]

The abundance of digital images has been facilitated by smartphones and inexpensive storage. Digital forensic investigation requires the processing of tons of digital images collected on devices to either identify or validate the device's user or to ascertain whether the operator has any connections to the case that would be of interest. Examining and evaluating heterogeneous activity presents several difficulties, including variability, complex interaction across information, and volume. Digital forensics processes are said to need the inspection and analysis stages. This research presents a hybrid optimization of the Grey Wolf and artificial bee colony (GW-ABC) optimization with deep learning model Convolutional Neural Network (CNN) i.e., GW-ABC-CNN, and the developed framework is integrated as a module for Autopsy software. The main objective of this research is to detect the heterogeneous activity of humans from the Heterogeneous Human Activity Recognition (HHAR) database. The developed model is integrated into the data-source ingest module; in this module, pre-processing, feature extraction, and detection process is performed. Moreover, in the pre-processing stage, the Min-Max normalization method is used and the required frequency and time features are extracted using the GW-ABC method. In addition, CNN is used to detect heterogeneous activity; this detection process is performed by four layers. Finally, the effectiveness of the developed model is assessed, and the outcomes of using the GW-ABC-CNN paradigm were compared to those of other strategies to evaluate the model's effectiveness.

Proximal‐distal differences in muscle activity are rarely considered when defining the activity level of hamstring muscles. The aim of this study was to determine the inter‐muscular and proximal‐distal electromyography (EMG) activity patterns of hamstring muscles during common hamstring exercises. Nineteen amateur athletes without a history of hamstring injury performed 9 exercises, while EMG activity was recorded along the biceps femoris long head (BFlh) and semitendinosus (ST) muscles using 15‐channel high‐density electromyography (HD‐EMG) electrodes. EMG activity levels normalized to those of a maximal voluntary isometric contraction (%MVIC) were determined for the eccentric and concentric phase of each exercise and compared between different muscles and regions (proximal, middle, distal) within each muscle. Straight‐knee bridge, upright hip extension, and leg curls exhibited the highest hamstrings activity in both the eccentric (40%‐54%MVIC) and concentric phases (69%‐85%MVIC). Hip extension was the only BF‐dominant exercise (Cohen's d = 0.28 (eccentric) and 0.33 (concentric)). Within ST, lower distal than middle/proximal activity was found in the bent‐knee bridge and leg curl exercises (d range = 0.53‐1.20), which was not evident in other exercises. BFlh also displayed large regional differences across exercises (d range = 0.00‐1.28). This study demonstrates that inter‐muscular and proximal‐distal activity patterns are exercise‐dependent, and in some exercises are affected by the contraction mode. Knowledge of activity levels and relative activity of hamstring muscles in different exercises may assist exercise selection in hamstring injury management. [ABSTRACT FROM AUTHOR]