AI For Healthcare Research Areas

Medical Multimodal LLMs & On-Premise Assistance

Develops efficient medical language and vision-language systems for clinical assistance, medical visual question answering, and private on-premise deployment in low-resource or hardware-constrained environments.

Paper 1: Advancing Healthcare in Low-Resource Environments Through an Optimization and Deployment Framework for Medical Multimodal Large Language Models

Paper 2: Democratizing MLLMs in Healthcare: TinyLLaVA-Med for Efficient Healthcare Diagnostics in Resource-Constrained Settings

Paper 3: MedAide: Leveraging Large Language Models for On-Premise Medical Assistance on Edge Devices

MS Lesion Evaluation & Brain Image Analysis

Develops clinically meaningful evaluation and segmentation methods for brain imaging, with emphasis on multiple-sclerosis lesions, structured bias analysis, contextual segmentation, and reliable assessment beyond overlap-only metrics.

Paper 1: Beyond Dice: Clinically Structured Bias Analysis of Multiple Sclerosis Lesion Segmentation Models

Paper 2: Rethinking Evaluation of Multiple Sclerosis (MS) Lesion Segmentation Models

Paper 3: CA-U-Net: Context Aware U-Net for Brain Tumor Segmentation

Embedded Medical Image Segmentation

Designs lightweight CNN and Vision Transformer frameworks for accurate medical image segmentation on embedded and resource-constrained hardware using reduced-resolution inputs and model-complexity optimization.

Paper 1: J-Net: A Low-Resolution Lightweight Neural Network for Semantic Segmentation in the Medical field for Embedded Deployment

Paper 2: Embedded-ViT: A Framework for Embedded Deployment of Vision-Transformer in Medical Applications

Paper 3: Embedded Deployment of Semantic Segmentation in Medicine through Low-Resolution Inputs

Weakly Supervised Medical Segmentation

Explores weakly supervised and ensemble-based segmentation methods that learn from image-level labels, refine object boundaries, and reduce the annotation burden for medical imaging datasets.

Paper 1: A Novel Weakly Supervised Semantic Segmentation Ensemble Framework for Medical Imaging

Paper 2: ReFit: A Framework for Refinement of Weakly Supervised Semantic Segmentation Using Object Border Fitting for Medical Images

Paper 3: BoundaryCAM: A Boundary-based Refinement Framework for Weakly Supervised Semantic Segmentation of Medical Images

Ultrasound Imaging & Beamforming

Develops resource-efficient neural beamforming and accelerator techniques for high-frame-rate ultrasound imaging, reducing computation and hardware cost while supporting real-time diagnostic acquisition.

Paper 1: FeatuReBeam: Efficient Beamforming for High-Frame Ultrasound Imaging using Adaptive Feature Reuse with (m,k)-Guarantee

Paper 2: CapsBeam: Accelerating Capsule Network based Beamformer for Ultrasound Non-Steered Plane Wave Imaging on Field Programmable Gate Array

Paper 3: Tiny-VBF: Resource-Efficient Vision Transformer based Lightweight Beamformer for Ultrasound Single-Angle Plane Wave Imaging

Develops resource-efficient neural beamforming and accelerator techniques for high-frame-rate ultrasound imaging, reducing computation and hardware cost while supporting real-time diagnostic acquisition.

Paper 1: Hy-Deft: A Hybrid Defense Technique for Vision Transformers against Adversarial Attacks in Medical Imaging

Paper 2: S-E Pipeline: A Vision Transformer (ViT) Based Resilient Classification Pipeline for Medical Imaging Against Adversarial Attacks

Robust & Secure Medical Imaging AI

© 2026 eBRAIN