Papers.Data.Code
Dataset

Datasets used in FocusMIL paper

Raymvp12

Contains pre-extracted 10x 256x256 patch features for Camelyon16 slides, plus a poisoned MIL benchmark and localization features for WSI classification.

Hugging Face ↗Jul 12, 2026Paper ↗Code ↗

  • cv
  • multiple-instance-learning
  • computational-pathology
  • camelyon16
Pipeline diagram showing ResNet18 encoder-classifier method for histopathology classification
Pipeline diagram showing ResNet18 encoder-classifier method for histopathology classification

What it is

Pre-extracted Camelyon16 patch features plus a shortcut-poisoned MIL test set for robust whole-slide classification and localization without raw WSIs. ~23 GB of pre-extracted Camelyon16 patch features across 4 feature sets: ResNet18, CTransPath, a semi-synthetic Standard MIL test, and FROC features with coords, plus 47 tumor masks.

Key results

  • 4 feature directories plus 47 GT tumor masks
  • ResNet18 512-d and CTransPath 768-d features
  • CC0-1.0 license; ~23 GB total

Useful for benchmarking robust MIL slide classification and FROC tumor localization without raw whole-slide images.

One useful ML email a week.

More from Jul 13 – Jul 19

The whole week →

Picked Jul 17

Picked Jul 16

Worth your time, not all of your time.

We track new ML work across papers, code, and data, then keep only the pieces worth a closer look. Selected, not collected.

Papers.Data.Code

The Weekly ML Signal

Selected, not collected.

One useful ML email a week: the papers, repos and datasets worth your attention.

Worth your time, not all of your time.