Add Space description, test-image pointer, and source link

Documents what the Space does, links back to the GitHub source repo,
and points testers at the TXL-PBC dataset for sample smear images.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TjcgYCZjyAPxP1vq8X9VTn

README.md

@@ -8,3 +8,19 @@ sdk_version: "4.0.0"
 app_file: "app.py"
 pinned: false
 ---
+
+# 🩸 SmartCBC — Multimodal AI Blood Smear Analysis
+
+Upload one or more peripheral blood smear FOV (field-of-view) images and get:
+
+- RBC / WBC / Platelet counts (YOLOv8 detector)
+- WBC subtype classification (ResNet50, 8 classes)
+- Aggregated multi-FOV differential
+- Age- and gender-specific reference-range comparison
+- Basic clinical insights (non-diagnostic)
+
+**Test images:** don't have a smear image handy? Sample FOVs are available from the public [TXL-PBC dataset](https://github.com/lugan113/TXL-PBC_Dataset) — any PNG from its `images/test` folder works.
+
+Source code: [SmartCBC-Multimodal-Pipeline](https://github.com/prashanth-ds-ml/SmartCBC-Multimodal-Pipeline) on GitHub.
+
+**Disclaimer:** this is a research prototype. Results are for qualitative screening only and are not a substitute for formal laboratory testing or clinical diagnosis.