Image Processing

Implementation of different image processing algorithms using python and matlab

1. Day and Night Classifier:

This project aims to classify day and night images using image processing.

We are going to use the HSV color system which isolates the value that varied under different lighting conditions.

Average brightness is high for day images.

We plotted the average brightness for all training images, we used ot to be above 95, and this a comparison between the tru lael and the estimated label of the image in the right.

We evaluated the classifier by the accuracy and it fails in 14 images out of 160 in the test images.
Accuracy: 0.9125
Number of misclassified images = 14 out of 160

2.Threshold Segementation:

This project aims to segment the green backgrounded car image and replace the its background by the sky background.

We used the Cv2 library and converted our image to an cv2 rgb image (image in the left), then we have applied our threshold selecting the green color in the background to creat a mask ( image in the center ) , then applied the mask to original image (image in the right).

we resized the background image as the same size of the car image, then applied the mask on it (image in the left), afterwards, we summed the masked image and the resized masked background (image in the right).

3.Texture Image Comparison:

Calculating a score based on the similarity of the two images using either cross-correlation, convolution or sum of squared differences, and comparing the implementation of these algorihms and the MLP classifier.

4.Gesture Detection:

Simple matlab code for webcam captured images (with white background) enhancement and segmentation then to recognize fingers count.