feat(phase1): implement OpenCV image processing pipeline

- Create ImageProcessor service with EXIF orientation detection
- Implement smart cropping via OpenCV contour detection (10% padding)
- Add text orientation detection using Hough line transform
- Resize and compress images to 1200px with 85% JPEG quality
- Generate 200px square thumbnails with center crop
- Fallback to Pillow if OpenCV fails
- Comprehensive test suite: 28 tests all passing
- File size validation (reject >10MB)
- Graceful error handling for corrupted/invalid images
- Update requirements.txt with opencv-python, piexif, python-magic

venv/lib/python3.12/site-packages/cv2/segmentation/__init__.pyi

@@ -0,0 +1,39 @@
+__all__: list[str] = []
+
+import cv2
+import cv2.typing
+import typing as _typing
+
+
+# Classes
+class IntelligentScissorsMB:
+    # Functions
+    def __init__(self) -> None: ...
+
+    def setWeights(self, weight_non_edge: float, weight_gradient_direction: float, weight_gradient_magnitude: float) -> IntelligentScissorsMB: ...
+
+    def setGradientMagnitudeMaxLimit(self, gradient_magnitude_threshold_max: float = ...) -> IntelligentScissorsMB: ...
+
+    def setEdgeFeatureZeroCrossingParameters(self, gradient_magnitude_min_value: float = ...) -> IntelligentScissorsMB: ...
+
+    def setEdgeFeatureCannyParameters(self, threshold1: float, threshold2: float, apertureSize: int = ..., L2gradient: bool = ...) -> IntelligentScissorsMB: ...
+
+    @_typing.overload
+    def applyImage(self, image: cv2.typing.MatLike) -> IntelligentScissorsMB: ...
+    @_typing.overload
+    def applyImage(self, image: cv2.UMat) -> IntelligentScissorsMB: ...
+
+    @_typing.overload
+    def applyImageFeatures(self, non_edge: cv2.typing.MatLike, gradient_direction: cv2.typing.MatLike, gradient_magnitude: cv2.typing.MatLike, image: cv2.typing.MatLike | None = ...) -> IntelligentScissorsMB: ...
+    @_typing.overload
+    def applyImageFeatures(self, non_edge: cv2.UMat, gradient_direction: cv2.UMat, gradient_magnitude: cv2.UMat, image: cv2.UMat | None = ...) -> IntelligentScissorsMB: ...
+
+    def buildMap(self, sourcePt: cv2.typing.Point) -> None: ...
+
+    @_typing.overload
+    def getContour(self, targetPt: cv2.typing.Point, contour: cv2.typing.MatLike | None = ..., backward: bool = ...) -> cv2.typing.MatLike: ...
+    @_typing.overload
+    def getContour(self, targetPt: cv2.typing.Point, contour: cv2.UMat | None = ..., backward: bool = ...) -> cv2.UMat: ...
+
+
+