Sciome · AI-Powered Science
From Raw Images to Biological Insight
A scalable, AI-driven platform for high-throughput biological imaging. Automated, reproducible, interpretable - from multi-well plates to actionable data.
20+
Morphological Endpoints
10
Organ Segmentations
2
Validated Modules
Fast
High-Throughput Ready
PixelVision is a scalable, AI-driven image analysis workbench designed to support high-throughput biological imaging data across multiple platforms, species, and experimental designs. It is particularly optimized for large-scale multi-well plate screening, enabling automated, reproducible phenotypic analysis across high-content, high-throughput studies. It serves as a unifying framework for organizing, visualizing, analyzing, and interpreting complex imaging data, bridging the gap between raw images and actionable biological insights.
Core Capabilities
What You Can Do
A unified framework built to scale from a single experiment to enterprise-level screening campaigns.
High-Throughput Screening
Automated phenotypic analysis across thousands of wells with consistent, reproducible scoring.
AI Morphology Detection
Deep learning models detect abnormalities, mortality, and structural changes at unprecedented scale.
Organ Segmentation
Precise organ- and cell-level quantification across multiple biological systems and imaging modalities.
QC & Human Review
Human-in-the-loop validation with image quality scoring, best-image selection, and override tools.
Metadata Integration
Unified experiment comparison with rich metadata linking dose, timepoint, and treatment conditions.
Microphysiological Systems (MPS)
Extend automated image analysis to MPS and organ-on-a-chip platforms - bringing the same reproducible, high-throughput scoring to complex in vitro models.
Benchmark Dose Modeling
Estimate BMD from dose-dependent morphological changes with built-in statistical workflows (coming soon).
Analysis Modules
Explore the Modules
Modules are the core functional building blocks of PixelVision. Each module is a self-contained, task-specific analysis pipeline designed to support a particular imaging platform, biological system, or experimental workflow.
The modules currently available are supported, production-ready implementations. However, PixelVision is inherently extensible:
- New modules can be added based on specific user needs.
- Existing modules can be modified or extended with new machine learning algorithms, endpoints, or workflows.
Currently, PixelVision includes two supported modules, all operating within a unified framework that ensures consistent data handling, visualization, and outputs while allowing the platform to scale and evolve with new assays, species, and imaging technologies.
FINS: Automated Zebrafish (Danio rerio) Image Analysis
Fish Imaging and Neural Scoring
Automates the full zebrafish imaging pipeline – from raw multi-well plate images to validated phenotypic calls – enabling large-scale developmental toxicology studies with speed and precision.
Morphology Scoring (20+ Endpoints)
- Axis curvature, scoliosis, body axis malformations
- Tail bending, fin absence, trunk abnormalities
- Craniofacial snout/jaw defects, edema, eye defects
- Notochord defects, yolk sac & pericardial edema
- Pigmentation abnormalities and otolith defects
Organ Segmentation — 10 Regions
- Eye, brain, otolith, swim bladder, notochord
- Yolk, pericardium, craniofacial, otic vesicles, pigment
Multi-View QC & Best-Image Selection
- Dorsal, lateral, and additional views supported
- Mortality at 24-hour and 5-day timepoints
- AI QC score auto-selects the best image per well
Frequently Asked Questions
What morphological endpoints does FINS detect?
FINS detects 20+ endpoints including axis curvature, scoliosis, tail bending, fin absence, trunk abnormalities, craniofacial defects, notochord defects, yolk sac edema, pericardial edema, pigmentation changes, and otolith defects in 5dpf Danio rerio larvae.
Which zebrafish organs can FINS segment?
FINS segments 10 regions: eye, brain, otolith, swim bladder, notochord, yolk, pericardium, craniofacial region, otic vesicles, and pigment – with quantitative area and morphology measurements per organ.
How does FINS handle multi-view images?
FINS processes dorsal, lateral, and additional views per well. AI-driven QC scoring ranks image quality and auto-selects the best image before running morphology analysis.
Is FINS validated for developmental toxicology?
Yes. Models are trained on large expert-annotated Danio rerio datasets and validated in peer-reviewed publications on PFAS developmental toxicity in SLAS Technology and Aquatic Toxicology.
CellVision: Cell Culture
Morphology Analysis
AI-Powered Cell Screening
CellVision brings intelligent morphological assessment to cell culture systems. Regional prediction maps reveal sub-well patterns invisible to manual review, enabling high-content screening at scale.
Classification & Prediction
- Normal, Stressed, and Abnormal category classification
- Interpretable regional prediction maps per well
- Supports 2D and 3D cell culture systems
Quantitative Analysis
- Quantitative morphological feature extraction
- Integrated QC scoring and validation pipelines
- Designed for scalable high-content screening workflows
Frequently Asked Questions
What cell categories does CellVision classify?
CellVision classifies wells into three categories: Normal, Stressed, and Abnormal — providing a nuanced readout beyond simple live/dead scoring.
What are the regional prediction maps?
Each well image is divided into a grid of sub-regions, each independently scored to create a spatial heatmap revealing edge effects or compound gradients invisible in well-level averages.
Does CellVision support 3D cell cultures?
Yes. CellVision works with both 2D monolayer and 3D systems including spheroids and organoids, making it suitable for physiologically relevant in vitro models.
How does CellVision integrate with existing workflows?
CellVision operates within the PixelVision framework, ingesting images from high-content imaging platforms. Results integrate with the platform’s QC, metadata, and BMD calculation tools (coming soon).
SpheroidVision: Spheroid & Organoid Analysis
3D Morphology Analysis & Growth Analysis
SpheroidVision enables automated, high-throughput analysis of 2D and 3D spheroids and organoids, capturing complex morphology, growth dynamics, and treatment responses with precision and scalability.
SEGMENTATION & STRUCTURE QUANTIFICATION
- Robust spheroid and organoid segmentation (brightfield + fluorescence)
- Multi-channel image support for integrated analysis across imaging modalities
- Cellular and nuclear segmentation within spheroids
- Integrated 3D visualization and exploration tools
MORPHOLOGICAL FEATURES
- Size, area, volume, and compactness
- Circularity, eccentricity, and boundary irregularity
- Texture-based heterogeneity and density metrics
GROWTH & KINETICS TRACKING
- Longitudinal spheroid growth curves
- Time-resolved treatment response profiling
- Viability assessment and shrinkage/expansion dynamics
- Fully compatible with time-lapse imaging workflows
Frequently Asked Questions
What types of models does SpheroidVision support?
SpheroidVision supports both 2D and 3D culture systems, including tumor spheroids, organoids, and co-culture models, enabling analysis of physiologically relevant in vitro systems.
What features does SpheroidVision quantify?
SpheroidVision delivers multi-scale morphological quantification, capturing features such as size, shape, and heterogeneity at both the whole-spheroid and individual cell levels.
Can SpheroidVision analyze spheroid growth over time?
Yes. SpheroidVision supports longitudinal analysis of spheroid growth, enabling time-resolved tracking of expansion, shrinkage, and treatment response across time-lapse experiments.
Does SpheroidVision support fluorescence imaging?
Yes. The platform supports both brightfield and fluorescence imaging, enabling multi-channel analysis such as tumor labeling, viability markers, and co-culture systems.
Transparent Pricing
Simple, Volume-Based Plans
Pay for what you use.
FREE PILOT Analyze up to 500 images at no cost – no commitment required.
Tier 1
Standard
Under 10k images
$0.10/img
Most Popular
Tier 2
Professional
10k-200k images
$0.08/img
Tier 3
Enterprise
200k+ images
$0.06/img
On-Premise
Hardware Development
Deploy at your facility
Let's Discuss
Tailored to your infrastucture
Need a custom module or integration? Talk to our team →
What Scientists Say
Trusted by Researchers
"The Sciome FINS automated machine learning tool represents a transformative advancement in high-throughput zebrafish screening in multi-well plates. By removing the significant bottleneck of manual image analysis, this easy to use platform dramatically accelerates data processing while maintaining accuracy and reproducibility. FINS enables my group to focus on biological insights rather than workflow constraints, significantly enhancing the pace of discovery in toxicology and chemical screening."
Dr. Robyn Leigh Tanguay
University Distinguished Professor, Environmental and Molecular Toxicology Director, Sinnhuber Aquatic Research Laboratorym Director, OSU/PNNL Superfund Research Center Oregon State University
Peer-Reviewed Research
Publications
Oktay A, McAfee A, Green A, Howard B, Truong L, Tanguay R, Shah R & Tandon A. (2025). “FINS: An Interactive Platform for Automated Zebrafish Image Analysis and Morphological Screening.” Submitted to SLAS Technology.