more hints

segmentation/hints.md

@@ -0,0 +1,53 @@
+### Steps of segmentation
+
+- remove bluring: `clear un-clear points (using Gaussian filter(low pass filter)) (kernal which has a rectangular of pixel which surrounded by pixel of image)`
+- thresholding: `Conversion an image from colour or grayscale into a binary image(make black and white)`
+- masking: `taking the region of interest only (ROI)`
+
+## Techniques of thresholding
+
+1. **Otsu’s segmentation**(Set the pixel to white; if they are greater than the threshold else, set it to black.)
+2. **Edge-based segmentation algorithms**(It highlights regions of high spatial frequency (excluded))
+3. **Clustering-based image segmentation algorithms**
+4. **Neural networks for segmentation** (CNN,Deep Learning) (Most efficient and common used)
+
+---
+
+### TODO:
+
+#### LPCS (License plate character segmentation )
+
+1. get-input (image of plate in any side-view)
+2. Character Detection
+    - Perspective Transform ( (adjust plate)
+    - thresholding image
+    - detect un-necessary dots
+    - character detection
+3. output results
+
+---
+
+### Problems
+
+1. image noise
+2. plate frame
+3. clinch ??
+4. rotation and illumination
+
+---
+
+### Algorithms:
+
+- Vertical protection : is based on binary image and good for image that are not heavily but is weak in recognition
+  between 2,3 persian digits. '٢' '٣'
+- Connected domain:is based on the connectivity of character for letters and numbers ,effect of segmentation are
+  perfect,persian character makes mistake with 'ب' ,'پ'
+- Template matching: is based on normalization which is to refine the characters into block containing no extra white
+  spaces(pixels) in all borders of the characters
+
+### Template matching:
+
+`is a method that detects the characters, comparing the extracted characters and the templates built in the database`
+> Template matching is performed after resizing the extracted character into the same size.
+
+> The different styles of characters and numbers are stored as a templates

segmentation/seg.py

@@ -2,3 +2,12 @@ class DetectionCharacter:
     def __init__(self, imgplate):
         self.licensePlate = imgplate
         print("Segmentation Constructor")
+
+    def removeBluring(self, img):
+        print("remove bluring func")
+
+    def thresholding(self, img):
+        print("threshold func")
+
+    def masking(self,img):
+        print("masking func")