CPR: Chained Perceptual Refinement for Coarse-to-Fine Medical Image Classification

2026-07-01

AI Summary

High-resolution medical images contain fine-grained yet spatially sparse diagnostic cues, but processing them at full resolution incurs prohibitive computational costs, while conventional downsampling often discards critical details. To address this, this work proposes the Chain-wise Perception Refinement (CPR) framework, which formulates image analysis as a sequential decision process progressing from global to local perspectives: it dynamically identifies regions requiring refinement based on a low-resolution overview, extracts high-resolution evidence from the original image, and incrementally integrates global context. By fixing the backbone input size and progressively contracting the perceptual field, CPR preserves diagnostically relevant details under constant peak GPU memory usage, achieving a synergistic balance between accuracy and efficiency. Experiments across five medical imaging datasets demonstrate consistent superiority over existing methods, with accuracy gains up to 2.27% and up to 19.6× fewer GFLOPs at equivalent performance levels.

Abstract

High resolution medical images contain fine grained, spatially sparse cues that are critical for diagnosis, yet preserving full resolution incurs substantial computational and memory costs. Most deep models process images uniformly, leading to redundant computation or loss of diagnostic detail under downsampling. We propose Chained Perceptual Refinement, CPR, a coarse to fine framework that formulates medical image analysis as a sequential global to local decision process. Starting from a low resolution global view, CPR dynamically predicts the location and spatial extent of refinement regions, extracts high resolution evidence from the original image, and incrementally integrates it with global context. By keeping the backbone input size fixed while contracting the perceptual field, CPR preserves diagnostic fidelity with constant peak GPU memory. Extensive experiments on five medical imaging datasets and multiple backbone architectures demonstrate that CPR consistently outperforms both fixed resolution and multi scale state of the art baselines, achieving improvements of up to 2.27 percentage points over the second best method. It also achieves up to a 19.6 fold reduction in GFLOPs at matched accuracy, establishing a superior accuracy and efficiency trade off for high resolution medical image analysis. The code is available on GitHub.

Problem

• medical image classification
• high resolution
• computational efficiency
• diagnostic detail
• coarse-to-fine

Innovation

• Chained Perceptual Refinement
• coarse-to-fine
• medical image classification
• dynamic region refinement
• computational efficiency