Prompt-Conditioned FiLM and Multi-Scale Fusion on MedSigLIP for Low-Dose CT Quality Assessment

TL;DR

This paper introduces a prompt-conditioned framework using FiLM and multi-scale pooling on MedSigLIP for low-dose CT image quality assessment, achieving state-of-the-art results with data-efficient learning.

Contribution

It presents a novel prompt-guided approach with FiLM and multi-scale fusion for improved low-dose CT quality evaluation, outperforming existing methods.

Findings

Achieved PLCC=0.9575, SROCC=0.9561, KROCC=0.8301 on LDCTIQA2023.

Surpassed top challenge submissions with a prompt-guided model.

Demonstrated data-efficient learning and rapid adaptation capabilities.

Abstract

We propose a prompt-conditioned framework built on MedSigLIP that injects textual priors via Feature-wise Linear Modulation (FiLM) and multi-scale pooling. Text prompts condition patch-token features on clinical intent, enabling data-efficient learning and rapid adaptation. The architecture combines global, local, and texture-aware pooling through separate regression heads fused by a lightweight MLP, trained with pairwise ranking loss. Evaluated on the LDCTIQA2023 (a public LDCT quality assessment challenge) with 1,000 training images, we achieve PLCC = 0.9575, SROCC = 0.9561, and KROCC = 0.8301, surpassing the top-ranked published challenge submissions and demonstrating the effectiveness of our prompt-guided approach.