# DL4CV team#10: Lethal Company segmentation — Data

来源：https://www.kaggle.com/competitions/dl4cv-team10-lethal-company-segmentation/data

采集时间：2026-10-18T14:07:00Z

## Files

*   **train/** - the training set, each screenshot (PNG) in `train/screenshots/` is labelled through a pair of binary masks (PNG),  that are stored in `train/masks_loot/` and `train/masks_enemies/`; mask filenames match screenshot filenames
*   **test/** - the test set, only screenshots, enemy and loot masks are for you to predict
*   **sample_submission.csv** - a sample submission file in the correct format

## Columns

*   `img_id` - file name, e.g `3631_0007875.png`
*   `enemy_rle` - RLE for the predicted enemy mask, must use encoding described/implemented in the starter notebook and input image must be 860(w)x520(h)
*   `loot_rle` - RLE for the predicted loot mask, must use encoding described/implemented in the starter notebook and input image must be 860(w)x520(h)