PathoJev: Efficient Pathology Decisions from Shared Visual Context

查看原文 / View Original 编辑 / Edit

Computer Science > Computer Vision and Pattern Recognition arXiv:2610.40752 (cs) [Submitted on 29 Oct 2026] Title: PathoJev: Efficient Pathology Decisions from Shared Visual Context Authors: Chengyang Zhang, Wenchuan Zhang, Bo Li, Xinyu Liu, Jiaming Yang, Mengran Li, Jie Chen, Wei Ju, Yuhao Yi, Hong Bu, Jiancheng Lv.

计算机科学 > 计算机视觉与模式识别 arXiv:2610.40752 (cs) [提交于 2026 年 10 月 29 日] 标题：PathoJev：基于共享视觉上下文的高效病理决策作者：Chengyang Zhang, Wenchuan Zhang, Bo Li, Xinyu Liu, Jiaming Yang, Mengran Li, Jie Chen, Wei Ju, Yuhao Yi, Hong Bu, Jiancheng Lv。

Abstract: Pathology workflows require multiple visual judgments about the same slide, from identifying tumor type to assessing grade and invasion. Generating and parsing a separate textual response for each structured question adds latency between visual evidence and usable decisions.

摘要：病理工作流程需要对同一张切片进行多次视觉判断，从识别肿瘤类型到评估分级和浸润程度。为每个结构化问题生成并解析单独的文本响应，增加了视觉证据与可用决策之间的延迟。

Building on the System One paradigm, we introduce PathoJev, a pathology model that directly predicts answer probabilities for whole-slide and region-level assessment. PathoJev encodes each slide once and evaluates independent question suffixes in batches over the shared visual context.

基于“系统一”（System One）范式，我们引入了 PathoJev，这是一种直接预测全切片和区域级评估答案概率的病理模型。PathoJev 对每张切片进行一次编码，并在共享的视觉上下文上批量评估独立的问题后缀。

Reusing the hybrid backbone’s recurrent states further limits the computation that grows with slide-context length for each additional question. Across three whole-slide benchmarks, the 9B and 27B models achieve mean accuracies of 77.2% and 78.1%, exceeding SlideChat by 5.4 and 6.3 percentage points.

通过重用混合骨干网络的循环状态，进一步限制了随每个额外问题而增加的切片上下文长度相关的计算量。在三个全切片基准测试中，9B 和 27B 模型实现了 77.2% 和 78.1% 的平均准确率，分别超过 SlideChat 5.4 和 6.3 个百分点。

Both scales also maintain competitive accuracy on five ROI benchmarks. On WSI-Bench, the 9B model runs 9.1x faster than SlideChat. Shared context accelerates eight-question workloads by 6.9x and 5.9x over separate prefix encoding, rising to 11.2x and 8.9x for 64 questions.

这两种规模的模型在五个 ROI（感兴趣区域）基准测试中也保持了具有竞争力的准确性。在 WSI-Bench 上，9B 模型的运行速度比 SlideChat 快 9.1 倍。与单独的前缀编码相比，共享上下文将八个问题的任务负载加速了 6.9 倍和 5.9 倍，在处理 64 个问题时加速比提升至 11.2 倍和 8.9 倍。

These results provide a reusable probabilistic interface for combining multiple visual judgments in whole-slide and local tissue assessment. Code: this https URL. Subjects: Computer Vision and Pattern Recognition (cs.CV).

这些结果提供了一种可重用的概率接口，用于结合全切片和局部组织评估中的多种视觉判断。代码：见此 https 链接。学科：计算机视觉与模式识别 (cs.CV)。