Scribd
Upload
74 views1 page

Assignment 4

This document outlines four assignments for a Digital Image Processing course with the following deadlines and requirements: 1. Design an algorithm to detect when bottles on a conveyor belt are not filled to the midway point between the bottom of the neck and shoulder, based on sample image "bottles.tif" and stating assumptions. 2. Write an algorithm using region or boundary descriptors to automatically detect which yoga pose ("asana") is represented in example images, explaining the algorithm and its limitations. 3. Add Gaussian noise to an image and implement a wavelet-based denoising algorithm like wavelet shrinkage to enhance the degraded image. 4. Implement the Ripple transform on an image.

Uploaded by

Kartik Dutta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
74 views1 page

Assignment 4

This document outlines four assignments for a Digital Image Processing course with the following deadlines and requirements: 1. Design an algorithm to detect when bottles on a conveyor belt are not filled to the midway point between the bottom of the neck and shoulder, based on sample image "bottles.tif" and stating assumptions. 2. Write an algorithm using region or boundary descriptors to automatically detect which yoga pose ("asana") is represented in example images, explaining the algorithm and its limitations. 3. Add Gaussian noise to an image and implement a wavelet-based denoising algorithm like wavelet shrinkage to enhance the degraded image. 4. Implement the Ripple transform on an image.

Uploaded by

Kartik Dutta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

ASSIGNMENT4

DIGITAL IMAGE PROCESSING (DIP) - CSE 478


DEADLINE: 9TH OCTOBER (FRIDAY)

(1) A company that bottles a variety of industrial chemicals has heard of your success solving imaging problems and hires you to design an approach for detecting
when bottles are not full. The bottles appear as shown in bottles.tif image, as
they move along a conveyor line past an automatic filling and capping station.
A bottle is considered imperfectly filled when the level of the liquid is below
the midway point between the bottom of the neck and shoulder of the bottle.
The shoulder is defined as the region of the bottle where the sides and slanted
portion of the bottle intersect. The bottles are moving, but the company has
an imaging system equipped with a illumination flash front end that effectively
stops motion, so you will be given images similar to bottles.tif. Based on the
material you have learned in the class, write a program for detecting bottles that
are not filled properly. State all the assumptions that you make and that are
likely to impact the solution you propose.

(2) Assume that you have an image capturing system which can grab silhouettes
given an image of yoga asanas. Four example images are shown in the folder yogasan (namely y1-Ustrasana, y2- veerbhadrasan, y3-vrikhsasana, y4- trikonasana).
Write an algorithm (and code it) using region or boundary descriptors to automatically detect which asana is represented in the image. Explain your algorithms and state its limitations (would your algorithm work, if images of more
asanas are added in the dataset?)

(3) Take an image of your choice and add gaussian noise to it. Implement a wavelet
based denoising algorithm to enhance the degraded image. [Note: One of the
simplest approaches is called wavelet shrinkage and the idea is to threshold all
high frequency bands (detail bands) using a fixed value and reconstruct back
the image. For more advanced solutions, read about SURESHRINK and VISUSHRINK algorithms]

(4) Implement Ripple transform on an image of your choice (Lecture 10, non planar
transforms).

You might also like