Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

Muscle Coordination and Action

Muscle Coordination and Action

Muscle coordination is a complex and finely tuned process essential for smooth and purposeful movements like flexion, extension, adduction, abduction, and rotation. The human body orchestrates the actions of various muscles working in concert, each with a specific role. Four functional types describe how muscles work together: agonist, antagonist, synergist, and fixator.
Agonists
Agonist muscles, often called prime movers, are the primary muscles responsible for producing a specific movement.

Planar Rigid-Body Motion

Planar Rigid-Body Motion

Understanding the movement of a rigid body in planar motion involves recognizing that every particle within this body is traversing a path that maintains a consistent distance from a specific plane. This concept is fundamental in the study of physics and mechanical engineering, and it allows us to comprehend better how objects move in space.
Planar motion is typically divided into three distinct categories. The first is rectilinear translation, demonstrated by a subway train that moves along...

Fixed Action Patterns

Fixed Action Patterns

A fixed action pattern (FAP) is a specific, hard-wired sequence of behaviors that occurs in response to an external stimulus, called a sign stimulus. The behavior is “fixed” because it is essentially unchangeable—proceeding similarly across individuals of a species every time it occurs.

Generation of Action Potential in Skeletal Muscles

Generation of Action Potential in Skeletal Muscles

Every cell in the body maintains a membrane potential due to an uneven distribution of positive and negative charges across its plasma membrane. The membrane potential is measured in millivolts and quantifies the difference in charge across the membrane.
Like neurons, muscle cells are also regarded as excitable due to their capacity to change in response to stimuli, primarily due to voltage-gated ion channels embedded in their plasma membranes, which get activated by alterations in the cell's...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

LLM-DWA: a hybrid path planning framework combining large language models with the dynamic window approach.

Scientific reports·2026

Same author

Lightweight Multimodal Fusion for Urban Tree Health and Ecosystem Services.

Sensors (Basel, Switzerland)·2026

Same author

Enhanced real-time 6d pose estimation for automatic recovery of in-flight UAVs using distance-aware keypoint heatmaps.

Scientific reports·2025

Same author

Enhancing Medical Image Segmentation and Classification Using a Fuzzy-Driven Method.

Sensors (Basel, Switzerland)·2025

Same author

3D-printed shadow masks for micro-patterned electrodes.

RSC advances·2024

Same author

Improving Fire Detection Accuracy through Enhanced Convolutional Neural Networks and Contour Techniques.

Sensors (Basel, Switzerland)·2024

Search research articles

Related Experiment Videos

SPARK: sparse-perception action recognition with keyframes for quadruped robots.

Sehun Park¹, Andrew Jaeyong Choi²

¹School of Computing, Gachon University, Seongnam, 13120, Republic of Korea.

Scientific Reports

|May 8, 2026

Summary

This summary is machine-generated.

This study presents a lightweight Human Action Recognition (HAR) model that efficiently processes video data by selecting keyframes. This approach significantly reduces computational costs for real-world robotic applications.

Keywords:

Human action recognition Quadruped robot Temporal transformer

Related Experiment Videos

Area of Science:

Computer Science
Robotics
Artificial Intelligence

Background:

Human Action Recognition (HAR) models traditionally face high computational costs due to processing extensive video data.
Efficient HAR is crucial for real-time applications, especially in resource-constrained environments like robotics.

Purpose of the Study:

To introduce a lightweight Human Action Recognition (HAR) model optimized for computational efficiency.
To reduce data redundancy by strategically selecting informative keyframes for processing.
To enable on-device action recognition in resource-constrained robotic systems.

Main Methods:

Utilized a pre-trained DaViT backbone for effective feature extraction.
Employed a Temporal Transformer to capture spatial and temporal dynamics from sparse keyframes.
Implemented a keyframe selection strategy to minimize data redundancy.

Main Results:

The proposed HAR model significantly reduces the computational cost compared to traditional 3D CNNs and LSTMs.
Demonstrated successful deployment on a quadruped robot for on-device action recognition.
Achieved an efficient inference pipeline for real-time HAR in robotic environments.

Conclusions:

The lightweight HAR model offers a computationally efficient solution for complex action recognition tasks.
The keyframe selection approach effectively addresses data redundancy challenges in video processing.
This work represents a significant advancement in applying HAR to resource-limited robotic platforms.