Image Analysis Platform
Nine Modules, Full Workflow Coverage
From image preprocessing to advanced phenotype exploration, from cell segmentation to spatial decoding — every step in one platform.
TR-AI — One Sentence Completes the Analysis
Say what you need; Assistant T understands it, plans the pipeline, and orchestrates every plugin — from preprocessing to spatial analysis.
TR-AI
No complex operation needed. Just say what you want — TR-AI plans the analysis, runs the right toolchain, and shows results in real time.
Auto-plans analysis, executes the full pipeline
Progress visible, immediate result preview
WeChat control — analysis at your fingertips
See Assistant T in action — request a demo link
TR-StainSync
Fully automatic alignment across multi-round staining, serial sections, and multi-modal pathology images — without manually picking a reference.
H&E / IHC / mIF and more
Comprehensive image format support
Composite registration engine
Whole-slide & TMA ready
TR-NoiseX
Smart denoising strips autofluorescence and non-specific background. Dynamic enhancement unifies visual and algorithmic reads. Metadata repair standardizes 100+ formats.
Removes autofluorescence, lifts downstream accuracy
Boosts channel contrast, unifies human & AI views
Format support + metadata repair
Channel trimming & dtype conversion
Fast merge into a standard multi-channel file
TR-TMA
Automatic TMA core detection eliminates tedious per-core adjustments. Fluorescence-ready identification corrects problems in native tools.
Auto core detection, fewer manual tweaks
Fixes mIF TMA detection instability
Convert to standard batch-analysis objects
Auto-detected core overlays
TR-Seg
Nucleus, membrane, and dense-cell segmentation modes — expert-grade identification precision that gets closer to biological reality.
Nucleus-level boundary detection
Membrane-level detection
Expert-level identification precision
Dense cell clusters segmented cleanly
TR-PhenoClass
Deep-learning phenotype classification that goes beyond mean-intensity thresholding. Distinguish tumor, immune, and stromal populations with reusable marker rules.
Marker-logic rules — auditable, re-verifiable
Mature panel rules apply to batch projects
Major classes first, subtypes after
Phenotype overlay on original image
TR-PhenoDiscover
Leiden clustering + UMAP visualization + image linkage — discover rare subtypes, then verify them interactively against the original image.
Find rare cell populations from high-D data
Cluster, matrix, and image linked in real time
Coarse cluster first, then drill into subsets
Interactive UMAP plots
TR-Spa
Quantify distance, neighborhood, and co-localization. Reconstruct tumor-immune-stromal microenvironments. All in a unified spatial stats toolbox.
Pair-wise shortest distance between cell types
KNN / Delaunay spatial graphs
Cluster + enrichment around target cells
Spatial density & hotspot maps
TR-HECellSage
For routine H&E images, automatically identify tumor, immune, stromal, and necrotic cells — boosting quantitative H&E analysis efficiency.
Cell type ID directly on H&E
Common scenarios — out-of-the-box
Morphology + spatial joint analysis
Lifts routine pathology efficiency
Try the platform on your own data
Free 14-day evaluation with sample images and full plugin access.