• Python 2.7

• numpy 1.15

• scipy 1.2.1

• matplotlib

• imageio 2.6.1

• imutils 0.5.3

• opencv 2.4.13

• PyQt 5.6

1. Clone repo

   • git clone https://github.com/bushi180/PythonProjects.git

2. Install the required libraries. You can use environment.yml with conda command.

   • conda env create -f environment.yml

3. Execute python script in each directory.

Process is accomplished with the method introduced in the paper Colorization using Optimization by Anat Levin, Dani Lischinski, and Yair Weiss. It is based on the premise that neighboring pixels in space-time that have similar intensities should have similar colors.

The modified PyQt Scribble interface allows a user to import a black and white image, apply color marks, and through colorization using optimization produce a fully colored image. There are options to draw new color marks on an image or import an image pre-colored with marks.

Process is accomplished with the method introduced in the paper Seam Carving for Content-Aware Image Resizing by Shai Avidan and Ariel Shamir. It is based on the premise that a seam is an optimal 8-connected path of pixels on a single image from top to bottom or left to right. By repeatedly carving out or inserting low cost seams, defined by an energy map, the aspect ratio of an image can be changed. The selection and order of seams protect the content of the image.

Used a window based correlation method to determine which pixels in a left stereo image correspond with pixels in a right stereo image. To compare windows I used the sum of squared difference. The disparity map shows the difference in pixel coordinates of similar features in the left and right stereo images.

Online Simultaneous Localization and Mapping (SLAM). Calculates the location of a bot and landmarks after a series of bot movements and measurements to landmarks. Results in the mapping of the environment. Online version only maintains the most recent position in bot's path, preventing the omega matrix from becoming too large.

• red triangle - actual bot position and direction

• blue dash line - actual bot path

• blue dots - SLAM calculated bot positions

• purple dots- SLAM calculated landmark positions

An informed search algorithm that calculates the shortest path found between two nodes or graphs. Used an Euclidean distance heuristic.

• green circle - start position

• magenta star - goal position

• black squares - walls

• blue X's - searched nodes

• red line - shortest path found

An optimal estimation algorithm that predicts states from uncertain measurements. Used to calculate the position of people in frame. Used OpenCV Histogram of Gradients (HoG) descriptors as measurements to update the filter.

• green rectangle - bounding box created from Kalman filter state after prediction and measurement update.