CHISON SonoAI: Smart Diagnosis, New Era

Amid the rapid convergence of artificial intelligence and healthcare, CHISON stands at the forefront of AI-powered ultrasound innovation. Leveraging world-leading AI algorithms, the company has independently developed intelligent solutions covering the entire ultrasound diagnostic and therapeutic workflow — from intelligent scanning, precise measurement, and assisted diagnosis to interventional treatment and structured reporting.

Driven by clinical needs, CHISON integrates large language models with deep learning algorithms, uniting resources from industry, academia, and research. The company has tackled key challenges such as large-model ultrasound training, intelligent image interpretation, and multimodal data fusion, building a fully autonomous and controllable intelligent ultrasound diagnostic ecosystem.

Powered by AI large models, CHISON offers comprehensive application-specific solutions in cardiology, obstetrics & gynecology, vascular imaging, thyroid & breast imaging, and point-of-care ultrasound (POCUS). These solutions enable full-chain upgrades in intelligent scanning, quantitative measurement, assisted diagnosis, and automated reporting, delivering enhanced accuracy, efficiency, and intelligence to clinicians worldwide.

Cardiac Ultrasound AI Solution

Auto EF

Strain and Strain Rate

SonoCardiac

Built on a multi-level information fusion algorithm framework and grounded in cardiac hemodynamics, this solution dynamically captures cardiac function and performs automated quantitative assessments. The proprietary CS-Net architecture integrates three-dimensional myocardial motion tracking with spatial deformation modeling to precisely characterize myocardial motion patterns, improving strain analysis sensitivity and reliability for early detection of subtle dysfunction. A deep learning–based intelligent diagnostic system employs high-precision image segmentation to accurately capture apical four and two-chamber views, enabling rapid anatomical localization and image quality grading. The system’s “global–microscopic–local” three-tier diagnostic approach significantly improves early detection of cardiac dysfunction, offering a powerful tool for cardiovascular disease screening.

Obstetric & Gynecologic Ultrasound AI Solution

SonoAI OB

SonoFollicle

VirtualFace

This solution combines four key innovations to support reproductive health and perinatal care across the entire patient journey:

l Automated follicle measurement for efficient monitoring in assisted reproduction.

l SonoOB 4.0 — a deep learning–driven obstetric system enabling automated measurement and quality control of 38 standard fetal views, improving screening accuracy and workflow efficiency.

l VirtualFace technology reconstructs detailed fetal facial features, aiding in congenital anomaly screening while fostering parental bonding.

l 3D printing integration converts precise fetal models into physical keepsakes, meeting personalized family needs.

Vascular Ultrasound AI Solution

Combining a multimodal transformer framework with adaptive optimization techniques, this system enhances vascular identification, hemodynamic assessment, and interventional guidance:

l Intelligent Doppler optimizes beam angle and sampling range using adaptive parameter tuning and expert knowledge bases, minimizing operator-dependent error and improving spectral analysis accuracy.

l SonoVF differentiates arteries from veins via multimodal data fusion and deep neural networks, optimizing PICC line path planning to improve placement success and reduce thrombosis or mechanical injury risk.

l SonoVessel applies edge-weighted adaptive algorithms to extract vascular wall structures for more accurate lumen diameter measurement and elasticity assessment, improving stenosis/occlusion detection.

l SonoCPI uses bidirectional cross-attention mechanisms for precise atherosclerotic plaque segmentation, enabling high-precision vascular stenosis and sclerosis risk evaluation.

Thyroid & Breast Ultrasound AI Solution

Utilizing transformer architecture and deep learning, this solution integrates multi-scale lesion detection, quantitative imaging feature analysis, and AI-guided intervention planning:

l Automatic lesion contouring upon image freeze.

l AI-powered BI-RADS/TI-RADS classification to assess malignancy risk.

l Enhanced needle path planning and real-time tracking for precise intervention.

This full-process intelligent system — from screening and classification to intervention — improves diagnostic accuracy, operational efficiency, and standardization in thyroid and breast imaging.

Point-of-Care Ultrasound (POCUS) AI Solution

This solution combines deep neural networks, generative adversarial networks, and multimodal data fusion for rapid bedside evaluation:

l Automatic measurement of pleural line thickness and B-line counts with dynamic comparative reporting for early pulmonary disease detection.

l High-precision identification of musculoskeletal and neural structures to support ultrasounsd-guided nerve blocks.

l Automated bladder volume calculation to reduce false oliguria diagnosis risks.

l Diaphragmatic motion assessment using time-series analysis and regression modeling for respiratory function evaluation.

l SuperNeedle+ and SonoNeedle technologies enhance needle visualization through reinforcement learning and image enhancement, ensuring accurate puncture guidance.

By integrating AI-based recognition, decision support, and real-time feedback, this system delivers efficient, precise, and responsive bedside ultrasound diagnostics.

In this new era of deep integration between medical AI and clinical practice, CHISON will continue to invest in innovation, expanding the boundaries of intelligent ultrasound in primary screening, teleconsultation, and specialized care. By closing the loop between technology, clinical application, and service, the company drives equitable access to precision medicine and leads the next wave of digital transformation in healthcare.