Compressibility Analysis for the differentiable shift-variant Filtered Backprojection Model

2025-01-20

AI Summary

To address the high computational overhead and excessive parameter count of differentiable shift-variant filtered back-projection (FBP) models in non-circular-trajectory cone-beam CT—hindering practical deployment—this paper proposes, for the first time, a model compression framework integrating principal component analysis (PCA) directly into the differentiable FBP reconstruction pipeline. Specifically, redundant weight layers are decomposed into a trainable eigenvector matrix, low-dimensional compressed weights, and a mean vector. The method reduces model parameters by 97.25%, accelerates training significantly, and preserves reconstruction accuracy. Its key innovation lies in the tight integration of unsupervised dimensionality reduction with a physics-informed differentiable reconstruction model: this preserves physical interpretability while substantially improving computational efficiency, thereby enhancing both practical applicability and scalability of the model in real-world non-circular-trajectory imaging scenarios.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Non-convex Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Data transparency and provenance

Abstract

The differentiable shift-variant filtered backprojection (FBP) model enables the reconstruction of cone-beam computed tomography (CBCT) data for any non-circular trajectories. This method employs deep learning technique to estimate the redundancy weights required for reconstruction, given knowledge of the specific trajectory at optimization time. However, computing the redundancy weight for each projection remains computationally intensive. This paper presents a novel approach to compress and optimize the differentiable shift-variant FBP model based on Principal Component Analysis (PCA). We apply PCA to the redundancy weights learned from sinusoidal trajectory projection data, revealing significant parameter redundancy in the original model. By integrating PCA directly into the differentiable shift-variant FBP reconstruction pipeline, we develop a method that decomposes the redundancy weight layer parameters into a trainable eigenvector matrix, compressed weights, and a mean vector. This innovative technique achieves a remarkable 97.25% reduction in trainable parameters without compromising reconstruction accuracy. As a result, our algorithm significantly decreases the complexity of the differentiable shift-variant FBP model and greatly improves training speed. These improvements make the model substantially more practical for real-world applications.

Problem

• 3D Image Reconstruction
• Filtered Back-Projection
• Computational Efficiency

Innovation

• Principal Component Analysis (PCA)
• Differential Filter Back-Projection (FBP) Model
• Training Speed Enhancement