Individual observation of same-color mice
Inspect identities, contours and points separately, using occlusion and identity-ambiguity flags to review crossing intervals.
XUANHENG · BEHAVIORAL RESEARCHDemo available
Dense Tracking Studio
Multi-mouse body-surface point tracking
Link individual identities, contours and body-surface points in single-view videos. Inspect each mouse's fine-scale movement and verify tracking quality and interaction candidates against the source video.
- Contours and body-surface points together
- Multi-mouse results organized by individual
- Quality flags for manual review
Enlarge imageRESEARCH APPLICATIONS
Look beyond position to fine-scale body-surface movement
Analysis inputs Single-view video, initial targets and analysis regions
Fine-scale body-surface movement
Choose point density for the research question and inspect motion through linked source and close-up views.
Review of interactions and behavior candidates
Use trajectories, distances and time windows to narrow candidate intervals, then annotate and confirm events against the video.
CORE CAPABILITIES
Analyze, review, and organize results
Contours and points tracked together
Connect target localization, contour propagation, identity association and point tracking so outputs remain linked to the same individual.
Multi-mouse results by individual
Supports 1–3 mice with separate contour, point and status views. Complex occlusions require review.
Selectable point density
Choose 32/64/96 points per mouse and inspect their movement in three linked views.
Tracking quality and anomaly flags
Record occlusion, visible fraction and identity ambiguity to locate intervals requiring review.
Rule-based candidates and review
Generate candidates from trajectories, inter-animal distances and time windows, then confirm them through replay and annotation.
Evaluation and result delivery
Evaluate precision and recall against manual annotations and export trajectories, images, data and PDF.
Software in action
Actual interfaces for parameter configuration, result filtering, and review; original screenshots retain their Chinese labels.
Video import and initial target localization
Feature example
Enlarge imageMulti-mouse tasks and point density
Feature example
Enlarge imageRESEARCH OUTPUTS
Results and research interpretation
Measured examples from Dense Tracking Studio, preserving original figures and quality information.
Source image, contours and subject close-up
Tracking example
Enlarge imageTwo-mouse contour and body-surface point tracking
Actual analysis videoVideo preview
ANALYSIS WORKFLOW
From data preparation to research outputs
- 01
Import video and specify targets
Set the analysis region, subject count and initial locations, and verify that targets match the research subjects.
- 02
Configure body-surface point tracking
Select point density and tracking parameters for the research objective, then submit the task.
- 03
Inspect each subject
Use linked source, contour/point and close-up views to check occlusion, identity and point quality.
- 04
Review and confirm behavior candidates
Combine rule-based candidates, source video and manual annotations to confirm events and evaluate results when needed.
- 05
Export trajectories and research material
Save quality flags and confirmed results, then export trajectory data, images and PDF.
DELIVERABLES
Your analysis outputs
- Individual contours and body-surface point trajectories
- Frame-level quality flags
- Rule-based behavior candidates
- Data, images and PDF
BEFORE ANALYSIS
What to confirm before you start
- Body-surface points describe two-dimensional surface motion; they are not anatomical joints or three-dimensional pose measurements.
- Behavior candidates are generated by trajectory and time-window rules and require manual confirmation. They are not predictions from a general behavior-classification model.
- The two-mouse images come from a tested two-second, 60-frame clip; the video is a separate 30-second clip. They do not establish performance for all lighting, occlusion conditions or long videos.
FREQUENTLY ASKED QUESTIONS
Before you begin
How do body-surface points differ from skeletal keypoints?
Body-surface points follow visible surface motion and do not automatically correspond to anatomical landmarks such as the nose or joints. Anatomical or three-dimensional pose studies require a separately verified measurement approach.
Can results from crossings or occlusions be used directly for statistics?
First inspect identity, visible fraction and occlusion flags. Support for 1–3 subjects does not imply reliable identities under every crossing or occlusion condition.
Are behavior candidates final behavior labels?
No. Rule-based candidates narrow the search interval and must be checked against the behavior definition and video. Manual annotations can be used to evaluate candidate performance.
Find the right approach for your research
Tell us your experiment, recording angle, and target measurements to arrange a Dense Tracking Studio demo and sample assessment.
For research use only. Figure captions describe their sources and scope. Examples do not guarantee accuracy, diagnostic suitability, or research conclusions.