跳到正文
研踪PaperTrail
← 研踪 PaperTrail

CVF · CVPR 2026:cvf:4a6069d119da2e81c413d0278c25079d · 记录 #1

CryoKRAQEN: Kernel-Regularized Annealing for Quantized Embedding Networks in Cryo-EM Heterogeneous Reconstruction

来源日期(仅到年份) 2026
来源记录日期 2026
雷达观测 2026-11-01 23:58
CVF · CVPR 2026 ↗PDF ↗

摘要

Heterogeneous reconstruction in cryo-electron microscopy (Cryo-EM) is fundamental for understanding macromolecular structural diversity, yet remains challenging due to extreme noise, continuous conformational changes, and ambiguous image-to-structure mappings. Existing neural approaches often rely on encoder--decoder pipelines or fixed codebooks, which can be computationally demanding or struggle with complex heterogeneity. We propose CryoKRAQEN, a decoder-only framework that integrates triplane implicit representations with kernel-guided latent assignment and quantized embeddings to improve stability and structural discrimination. The method avoids encoder dependencies and mitigates collapse during training, enabling accurate modeling of both conformational and compositional variations. Across diverse Cryo-EM benchmarks, CryoKRAQEN delivers competitive performance, robust reconstructions, and interpretable latent organization compared to state-of-the-art neural and classical methods.

论文发现出处

这些页面提供论文线索,不代表同行评审或质量认证。

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

发表信息与关联版本

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

平台记录号代表本站观测到的元数据变化,不是作者修订号。来源日期仅到月份或年份时,按该月或该年首日存档。

记录 #1 · 2026 · 最新观测版本

CryoKRAQEN: Kernel-Regularized Annealing for Quantized Embedding Networks in Cryo-EM Heterogeneous Reconstruction

Heterogeneous reconstruction in cryo-electron microscopy (Cryo-EM) is fundamental for understanding macromolecular structural diversity, yet remains challenging due to extreme noise, continuous conformational changes, and ambiguous image-to-structure mappings. Existing neural approaches often rely on encoder--decoder pipelines or fixed codebooks, which can be computationally demanding or struggle with complex heterogeneity. We propose CryoKRAQEN, a decoder-only framework that integrates triplane implicit representations with kernel-guided latent assignment and quantized embeddings to improve stability and structural discrimination. The method avoids encoder dependencies and mitigates collapse during training, enabling accurate modeling of both conformational and compositional variations. Across diverse Cryo-EM benchmarks, CryoKRAQEN delivers competitive performance, robust reconstructions, and interpretable latent organization compared to state-of-the-art neural and classical methods.

查看此版本解读 →查看此版本原文 ↗