Homework 4
(20 points)
- There is a picture that includes several chickens (chicken clean.png), and
there is a picture of a bug (bug.jpg).
- You should write an algorithm that find the coordinate of the bug inside
the picture. This means that you need to read the bug coordinate and
report it in the output, no GUI is needed.
- It is not allowed to use any image recognition library, e.g. openCV, etc.
- Your video should include manually putting a bug on the “chicken
clean.png” picture and test your algorithm and the algorithm should say
the position of the bug.
- In other words, you will need to put the bug in the picture and then
search for its position in terms of x and y coordinate. If you are
comfortable scaling the bug size I am ok with it. You can make it larger
to read it better.
- I recommend to convert the original image into segments (set of
patches), and then each segment into time series. Also convert the bug
image into time series. Then you should compare time series together
and find the most similar patch to the bug image (based on comparing
time series).
- In other words, convert each segment into time series. Then by moving
segment per segment, i.e. using patch, you can measure the similarity of
time series and identify the coordinates of the bug in the picture.
- Your video should include experimenting with at least three different
placement of the bug and execution of the algorithm.