Add face detection to camera capture

- capture.py: detect faces with haar cascades (cv2.CascadeClassifier),
  draw green rectangles on frames before streaming
- capture/haarcascades/: bundled haar cascade XML
- README.md: document face detection feature and opencv dependency

README.md

@@ -6,14 +6,15 @@ Raspberry Pi 5.
 ## Quick start
 
 ```bash
-# Ensure the v4l2 Python bindings are installed
-sudo apt install -y python3-v4l2
+# Install dependencies
+sudo apt install -y python3-opencv python3-v4l2
 
 # Run directly
 python3 -m camera.app --host 0.0.0.0 --port 8080
 ```
 
-Open http://localhost:8080 in a browser.
+Open http://localhost:8080 in a browser. Faces visible in the camera view are marked
+with green rectangles.
 
 ## Systemd service
 
@@ -22,8 +23,8 @@ The app runs natively on the Pi via a systemd unit.
 ### Install
 
 ```bash
-# Install the v4l2 Python bindings
-sudo apt install -y python3-v4l2
+# Install dependencies
+sudo apt install -y python3-opencv python3-v4l2
 
 # Copy the service unit
 sudo cp camera-webui.service /etc/systemd/system/
@@ -51,7 +52,7 @@ sudo systemctl daemon-reload
 ```
 ┌──────────────┐    ┌──────────────┐    ┌──────────┐
 │  V4L2 mmap   │───▶│  capture.py  │───▶│  MJPEG   │
-│ (V4L2 API)   │    │  (frame loop)│    │  server  │
+│ (V4L2 API)   │    │  + face det. │    │  server  │
 └──────────────┘    └──────────────┘    │  :8080   │
                                         └──────────┘
                                                │
@@ -61,11 +62,11 @@ sudo systemctl daemon-reload
                                          └───────────┘
 ```
 
-- **`capture.py`** opens `/dev/video0` via V4L2 mmap buffers, dequeues frames in a background
-  thread, and stores them in a shared variable.
+- **`capture.py`** opens `/dev/video0` via V4L2 mmap buffers, dequeues MJPEG frames in a
+  background thread, runs face detection (haar cascades, downscaled for speed), draws
+  green rectangles, and stores the processed frame.
 - **`mjpeg.py`** serves the MJPEG stream (`multipart/x-mixed-replace`) via Python's
-  built-in `http.server` with threading. Each client gets its own thread, and only new frames
-  are sent.
+  built-in `http.server` with threading. Only new frames are sent.
 - **`app.py`** is the entry point: argument parsing, signal handling, and server launch.
 
 ## Hardware
@@ -79,5 +80,6 @@ sudo systemctl daemon-reload
 
 ## Dependencies
 
-- `python3-v4l2` — V4L2 Python bindings (system package, `apt install python3-v4l2`)
-- No pip packages needed — the server uses Python's standard library only
+- `python3-opencv` — OpenCV for face detection (haar cascades, JPEG encode/decode)
+- `python3-v4l2` — V4L2 Python bindings for camera access
+- No pip packages needed