Multi-camera Sync Capture
Multi-channel visible-light synchronized recording with one-click capture — time-aligned multi-view video for single mouse (4–6 cameras) and single pig (9-view) setups.
BehaviorAtlas MultiBehavior · Software Suite
Multi-behavior Analysis · Software Suite
Fine multi-behavior quantification for single subjects — markerless 3D capture, supervised behavior recognition, and model-driven behavior spectrum analysis, with dedicated hardware, software, and algorithms for single mouse and single pig.
Analysis Subjects
Built on a unified multi-behavior analysis methodology — dedicated hardware, software platforms, and algorithms for single mouse and single pig. Select a subject to view its technical configuration.
Software Functions
BehaviorAtlas Capture and BehaviorScope — multi-camera sync capture, TTL event sync, scene calibration, project management, 3D reconstruction, supervised behavior recognition, and export; configurable for single mouse and single pig.
TTL Event Sync
TTL event markers and external trigger sync — precise alignment of behavior event timestamps with multi-camera video for downstream analysis.
Auto-record After Acclimation
Set a mouse acclimation period — multi-camera recording starts automatically when it ends, reducing manual intervention and timing gaps.
Recording & Metadata
Experiment info, sample naming, and recording parameters — structured output to the project data folder as the foundation for tracking, reconstruction, and analysis.
Multi-camera Calibration
Checkerboard calibration for intrinsics/extrinsics, multi-camera alignment, and behavior event timestamps for accurate 3D fusion.
Supervised Behavior Recognition
Pre-trained models and supervised clustering identify and quantify behavior modules per model configuration — Supervised_paras segment and label outputs for fine multi-behavior analysis.
3D Skeleton Reconstruction
Fuses multi-view 2D tracking with calibration to reconstruct frame-by-frame 3D skeletons — csv and c3d coordinate outputs for downstream behavior recognition.
Built-in Default Models
Standard body-tracking and behavior-recognition models built in — follow the standard experimental workflow with no extra annotation, training, or custom model development in most cases.
Project & Data Management
Create and manage analysis projects with structured folders — data, 3Dskeleton, BeAOutputs, and results_export for traceable sample and output management.
2D Keypoint Tracking
Pre-trained networks detect and track body keypoints in single-view video, outputting per-frame 2D trajectories for 3D reconstruction.
Kinematic Parameter Computation
Automatic frame-level (FrameLevel_paras) and advanced kinematic parameters (AdvancedParameters) from 3D skeletons and behavior sequences.
Sync Replay & QC
Synchronously compare multi-camera video with 3D skeleton reconstruction — frame-by-frame QC for occlusion, dropouts, and calibration issues.
Visualization & Export
Behavior spectra, timelines, proportions, and group comparisons — one-click export of tables, figures, and videos to results_export.
Software Matrix
Capture and BehaviorScope — acquisition plus data analysis and visualization. Select a subject to explore its dedicated stack.
Capture
4–6 cams
Camera sync
50×50
cm arena
Multi-channel visible-light sync recording with TTL events and scene calibration — one-click recording.
BehaviorAtlas Capture · Single Mouse
Multi-camera acquisition software
Analysis
16+
Body keypoints
Supervised
Behavior recognition
AI keypoint detection, 3D skeleton reconstruction, and model-driven behavior spectra via supervised clustering — multi-dimensional visualization and one-click export.
BehaviorAtlas BehaviorScope · Single Mouse
Data analysis & visualization software
Analysis Workflow
Single-mouse workflow — calibration and capture through tracking QA, then batch analysis or a PoseLabeler custom-model path.
01
Camera Calibration
Checkerboard calibration for intrinsics/extrinsics and multi-camera alignment — the geometric foundation for 3D reconstruction.
02
Multi-camera Sync Capture
BehaviorAtlas Capture — multi-channel visible-light sync recording with TTL markers and experiment metadata.
03
Import into Analysis Software
Import captured video, calibration files, and experiment metadata into a BehaviorScope project for unified analysis.
04
Tracking Model Test
Test preset or custom tracking models on sample clips — evaluate keypoint accuracy and stability before full-batch analysis.
05
Batch Per-view Body Tracking
Batch body-keypoint tracking per camera view — high-precision 2D trajectories as input for 3D reconstruction.
06
3D Reconstruction
Multi-view fusion reconstructs 3D skeleton sequences for frame-by-frame pose data and downstream behavior analysis.
07
Behavior Recognition
Run built-in or custom behavior recognition models to extract and quantify behavior modules with labeled segments.
08
Results Viewing & Visualization
Review recognized behavior modules — explore spectra, timelines, statistical charts, and export publication-ready outputs.
When behavior recognition results are unsatisfactory
Annotate, train a custom model, then re-run recognition.
Behavior Annotation & Training
Annotate behavior segments in BehaviorScope and train a custom recognition model for your paradigm and phenotypes.
When tracking test results are unsatisfactory
If preset or custom tracking models show poor accuracy or stability on samples, annotate → train → import, then return to Tracking Model Test to re-validate.
PoseLabeler Body-point Annotation
Use our free PoseLabeler for efficient frame-level body-keypoint annotation and quality review.
Tracking Model Training
Train a custom body-keypoint tracking model from annotations — tailored to species, views, and setup.
Import Model into BehaviorScope
Import the trained tracking model into BehaviorScope, then return to the tracking model test step to re-validate before continuing.
Analysis Results
Single mouse — built-in default behavior models, frame-by-frame QC, data export, and visualization for standardized multi-behavior experiments.
01
Default Behavior Recognition Model
The built-in default model covers common mouse behaviors including moving forward, left turning, right turning, jumping, climbing, rearing, hunching, self-grooming, immobility, and sniffing — use it directly when your study focuses on these modules.
02
Data Review & Frame-by-frame QC
Review data frame by frame — kinematic parameters, 3D reconstruction, body-point tracking, and behavior recognition — to assess quality and flag anomalous segments.
03
Behavior Data Export
Export behavior recognition results and related parameters for custom statistics, group comparisons, and downstream analysis inside or outside the software.
04
Visualization & Chart Export
Built-in visualization for behavior spectra, timelines, transition matrices, and more — analyze and export publication-ready figures and reports in one click.
Get a Product Demo
Book an online or on-site demo — experience the full workflow with tailored protocol advice.
Book a Demo