[Alpenj]

What changed

• Added a local webcam dataset collector for four ImageFolder classes with a 1,000-image target per class.
• Added blur and near-duplicate filtering, resumable class counts, class renaming, and a localhost-only Python save API.
• Added an integrity checker and unit tests.
• Added a 17 KiB cafe-sharing ZIP containing only source, README, and tests.
• Linked the project from the repository root README.

Why

This provides the completed Step 1 data-labeling tool for the ResNet/PyTorch exercise while keeping personal webcam captures and machine-specific data out of Git.

User impact

Users can collect balanced four-class datasets locally, validate ImageFolder compatibility, and share the tool without uploading captured images.

Validation

• node --check app.js
• Python py_compile for server and integrity scripts
• python -m unittest -v test_integrity_check.py — 2 tests passed
• unzip -t ResNet_Dataset_Studio_STEP1.zip — no errors
• Privacy scan: no personal absolute paths, credentials, or captured images