Image object in mltu/annotations and code refactoring to work with it

Author: pythonlessons

CHANGELOG.md

@@ -1,4 +1,12 @@
-## [1.0.2] - 2022-03-... (unreleased)
+## [1.0.3] - 2022-03-20
+### Changed
+- Changed `mltu.augmentors` to work only with `Image` objects
+
+### Added
+- Created `Image` object in `mltu.annotations.image` to handle image annotations
+
+
+## [1.0.2] - 2022-03-20
 ### Changed
 - changes `OnnxInferenceModel` in `mltu.torch.inferenceModels` to load custom metadata from saved ONNX model
 - improved `mltu.dataProvider` to remove bad samples from dataset on epoch end
@@ -21,7 +29,7 @@
 - added `mltu.torch.models`, used to create PyTorch models, that wrapps whole model pipeline (training, validation, metrics, callbacks, etc.)
 - added `mltu.torch.callbacks`, used to create PyTorch callbacks, that may be used in training and validation
 - added `mltu.torch.metrics`, used to create PyTorch metrics, that may be used in training and validation
-- added 07_pytorch_tutorial tutorial
+- added `07_pytorch_tutorial` tutorial
 
 
 ## [1.0.0] - 2022-03-06

Tutorials/08_handwriting_recognition_torch/train_torch.py

@@ -80,7 +80,7 @@ def download_and_unzip(url, extract_to='Datasets', chunk_size=1024*1024):
     batch_size=configs.batch_size,
     data_preprocessors=[ImageReader()],
     transformers=[
-        # ImageShowCV2(), # uncomment to show images during training
+        # ImageShowCV2(), # uncomment to show images when iterating over the data provider
         ImageResizer(configs.width, configs.height, keep_aspect_ratio=False),
         LabelIndexer(configs.vocab),
         LabelPadding(max_word_length=configs.max_text_length, padding_value=len(configs.vocab))

Tutorials/README.md

@@ -5,4 +5,5 @@
 4. [Handwritten sentence recognition with TensorFlow](https://pylessons.com/handwritten-sentence-recognition), code in ```Tutorials\04_sentence_recognition``` folder;
 5. [Introduction to speech recognition with TensorFlow](https://pylessons.com/speech-recognition), code in ```Tutorials\05_speech_recognition``` folder;
 6. [Introduction to PyTorch in a practical way](https://pylessons.com/pytorch-introduction), code in ```Tutorials\06_pytorch_introduction``` folder;
-7. [Using custom wrapper to simplify PyTorch models training pipeline](https://pylessons.com/pytorch-introduction), code in ```Tutorials\07_pytorch_wrapper``` folder;
+7. [Using custom wrapper to simplify PyTorch models training pipeline](https://pylessons.com/pytorch-introduction), code in ```Tutorials\07_pytorch_wrapper``` folder;
+8. [Handwriting words recognition with PyTorch](https://pylessons.com/handwriting-recognition-pytorch), code in ```Tutorials\08_handwriting_recognition_torch``` folder;

mltu/__init__.py

@@ -1 +1,3 @@
-__version__ = "1.0.2"
+__version__ = "1.0.3"
+
+from .annotations.image import Image

mltu/annotations/__init__.py

@@ -0,0 +1,108 @@
+import os
+import cv2
+import typing
+import numpy as np
+
+class Image:
+    """ Image class for storing image data and metadata
+
+    Args:
+        image (str or np.ndarray): Path to image or numpy.ndarray
+        method (int, optional): OpenCV method for reading image. Defaults to cv2.IMREAD_COLOR.
+        path (str, optional): Path to image. Defaults to "".
+        color (str, optional): Color format of image. Defaults to "".
+    """
+    def __init__(
+        self,
+        image: typing.Union[str, np.ndarray],
+        method: int = cv2.IMREAD_COLOR,
+        path: str = "",
+        color: str = ""
+        ) -> None:
+        
+        if isinstance(image, str):
+            if not os.path.exists(image):
+                raise FileNotFoundError(f"Image {image} not found.")
+
+            self.image = cv2.imread(image, method)
+            self.path = image
+            self.color = "BGR"
+
+        elif isinstance(image, np.ndarray):
+            self.image = image
+            self.path = path
+            self.color = color
+
+        else:
+            raise TypeError(f"Image must be either path to image or numpy.ndarray, not {type(image)}")
+
+        self.method = method
+
+        if self.image is None:
+            return None
+
+        # save width, height and channels
+        self.width = self.image.shape[1]
+        self.height = self.image.shape[0]
+        self.channels = 1 if len(self.image.shape) == 2 else self.image.shape[2]
+
+    @property
+    def shape(self) -> tuple:
+        return self.image.shape
+
+    @property
+    def center(self) -> tuple:
+        return (self.width // 2, self.height // 2)
+
+    def RGB(self) -> np.ndarray:
+        if self.color == "RGB":
+            return self.image
+        elif self.color == "BGR":
+            return cv2.cvtColor(self.image, cv2.COLOR_BGR2RGB)
+        else:
+            raise ValueError(f"Unknown color format {self.color}")
+        
+    def HSV(self) -> np.ndarray:
+        if self.color == "BGR":
+            return cv2.cvtColor(self.image, cv2.COLOR_BGR2HSV)
+        elif self.color == "RGB":
+            return cv2.cvtColor(self.image, cv2.COLOR_RGB2HSV)
+        else:
+            raise ValueError(f"Unknown color format {self.color}")
+
+    def update(self, image: np.ndarray):
+        if isinstance(image, np.ndarray):
+            self.image = image
+
+            # save width, height and channels
+            self.width = self.image.shape[1]
+            self.height = self.image.shape[0]
+            self.channels = 1 if len(self.image.shape) == 2 else self.image.shape[2]
+
+            return self
+
+        else:
+            raise TypeError(f"image must be numpy.ndarray, not {type(image)}")
+
+    def flip(self, axis: int = 0):
+        """ Flip image along x or y axis
+
+        Args:
+            axis (int, optional): Axis along which image will be flipped. Defaults to 0.
+
+        Returns:
+            Object with flipped points
+        """
+        # axis must be either 0 or 1
+        if axis not in [0, 1]:
+            raise ValueError(f"axis must be either 0 or 1, not {axis}")
+
+        self.image = self.image[:, ::-1] if axis == 0 else self.image[::-1]
+
+        return self
+
+    def numpy(self) -> np.ndarray:
+        return self.image
+    
+    def __call__(self) -> np.ndarray:
+        return self.image