Computer Vision

A. Short answer type questions.

1. What do you mean by resolution?
Ans: The number of pixels in an image is called its resolution. It determines the level of detail or
clarity of images displayed on screens or captured by cameras.
2. Define CNN.
Ans: A Convolutional Neural Network is a deep Learning algorithm that is commonly used in image
recognition and processing. CNN analyses an image, extracts the best features and reduces its
size to make it manageable, while still preserving its important features. This helps it to
differentiate one image from the other.
3. What is meant by image convolution?
Ans: An image convolution is simply an element-wise multiplication of the image array with another
array called the kernel followed by sum. This results in forming a convolution matrix or filtered
image.
4. What are considered good features of an image during image processing?
Ans: The features like corners are easy to find as their exact location can be pinpointed in the
image. Thus, corners are always good features to extract from an image followed by the edges.
5. What are pixels?
Ans: The word "pixel" stands for "picture element". Every digital photograph is made up of tiny
elements called pixels. A pixel is the smallest unit of information that makes up a text, image or
video on a computer. Even a small image can contain millions of pixels of different colours. Pixels
are usually arranged in a 2-dimensional grid and are often in round or square shape.
6. State the objective of Computer Vision.
Ans: The objective of computer vision is to replicate both the way humans see and the way
humans make sense of what they see.
7. How are Computer Vision algorithms trained to identify image content?
Ans: Computer vision models are trained on massive amounts of visual data. Once a large amount
of data is fed through the model, the computer will "look" at the data and teach itself to differentiate
one image from other using deep learning algorithms.
8. Explain the concept of feature extraction in Computer Vision.
Ans: In image processing, the image can have features like a blob, an edge or a corner. These
features help us to perform certain tasks and analysis. Feature extraction refers to the process of
automatically extracting relevant and meaningful features from raw input images. The features like
corners are easy to find as their exact location can be pinpointed in the image, whereas the
patches that are spread over a line or an edge look the same all along.
B. Long answer type questions.
1. How do computers store RGB images?
Ans: Every RGB image is stored on a computer in the form of three different channels – the Red
channel, the Green channel and the Blue channel. Each channel contains a number of pixels, with
the value of each pixel ranging from 0 to 255. When all channels are combined together, they form
a coloured image. This means that in an RGB image, each pixel has a set of three different values
which together give colour to that particular pixel.
2. Explain CV and provide any two applications of it.
Ans: The Computer Vision domain of artificial intelligence enables machines to interpret visual
data, process it and analyse it using algorithms and methods to interpret real-world phenomena. It
helps machines derive meaningful information from digital images, videos and other visual inputs
and take actions based on that information.
Applications of Computer Vision are:
Face filters: This is one of the popular applications used in apps like Instagram and Snapchat. A
face filter is a filter applied to photographs, or videos in real time, to make the face look more
attractive. You can also use it to combine a face with animal features to give it a funny appearance.
Facial recognition: With smart homes becoming more popular, computer vision is being used for
making homes more secure. Computer Vision facial recognition is used to verify the identity of the
visitors and guests and to maintain a log of the visitors. This technology is also used in social
networking applications for detecting faces and tagging friends.
3. What is meant by the pixel value of a digital image?
Ans: Each pixel in a digital image on a computer has a pixel value which determines its brightness
or colour. The most common pixel format is the byte image, where this value is stored as an 8-bit
integer having a range of possible values from 0 to 255. Typically, zero is considered as no colour
or black and 255 is considered to be full colour or white.
4. Describe the CV tasks for a single object in an image.
Ans: The CV tasks for a single object in an image are:
Image Classification: This task involves assigning a label to the entire image based on its content.
Image Classification plus Localisation: This is the task which involves both processes of identifying
what object is present in the image and at the same time identifying at what location that object is
present in that image.
5. State the difference between CV and human vision.
Ans: Humans see an image with the help of their eyes, and then the brain processes and identifies
the image through learning and experience. In computer vision, AI first perceives the image with a
sensing device, and then computer vision and other AI algorithms identify and classify the
elements in the image to recognise it.
6. The Face Lock feature of a smart phone is an example of Computer Vision. Briefly discuss
localization? Give an example to illustrate this difference.
Ans: The face-lock feature on smartphones uses computer vision to analyse and identify facial
features. When a user activates the face-lock feature, the smartphone's CV system compares the
facial features with pre-registered photographs stored on the device. If the facial characteristics
match, the device grants access2 to the user. This authentication method offers convenience and
security for the user.
7. What is the difference between ‘image classification’ and ‘image classification and localization’.
Give an example to illustrate this difference.
Ans: Image classification involves identifying the main object category in a photo, while image
classification with localisation determines both the object's category and its precise location within
the image, often by drawing a bounding box around it.
For example, in an image showing a cat, the image classification algorithm will identify and label
the image as a cat. Whereas, the image classification with localization algorithm will not only
identify the cat, but will also draw a box to indicate the location of the cat in the image.
8. Explain the role and function of the convolution operator is CNNs
Ans: The convolution operator is a fundamental mathematical operation used in image processing.
It involves multiplying two arrays of numbers, element-wise, to produce a third array. In image
processing, convolution is a common tool used for image editing to apply filters or effects, such as
blurring, sharpening, outlining or embossing on an image.
An image convolution is simply an element-wise multiplication of the image array with another
array called the kernel followed by sum. This results in forming a convolution matrix or filtered
image.
9. List the different types of pooling that can be performed on an image in a CNN.
Ans: There are two types of pooling which can be performed on an image.
They are:
a. Max Pooling: This returns the maximum value from the portion of the image covered by the
Kernel.
b. Average Pooling: This returns the average value from the portion of the image covered by the
Kernel.
10. Apart from text- based search, search engines also carry out image-based searches. How
does Computer Vision help in this task?
Ans: Visual search algorithms in search engines use computer vision technology to help you
search for different objects using real world images. CV compares different features of the input
image to its database of images, analyses the image features and gives us the search result.
Computer vision, combined with machine learning allows the device not only to see the image, but
also to interpret what is in the picture, helping make decisions based on it.
11. How is Computer Vision used in the following fields?
Ans: Role of computer vision in the following fields:
a. Healthcare: Medical imaging has greatly benefited from computer vision. It not only creates and
analyses images, but also acts as an assistant and helps doctors to better understand a patient’s
health condition. CV analyses X-Rays, CT scans and MRI is used to read and convert 2D scanned
images into interactive 3D models. This results in an increase in the accuracy and efficiency of
diagnosis since the machines can identify details invisible to the human eye.
b. Warehouses: CV can be used in warehouses to remove human error during the receiving and
storing process of products by automating the scanning and data entry process for inventory
management. Robots equipped with CV accurately pick parcels and pack them. CV also
automatically checks the order against the contents.
12. “Computer Vision is the basis of autonomous vehicles.” Justify.
Ans: Autonomous driving involves identifying objects, getting navigational routes and monitoring
the surroundings. Automated cars from companies like Tesla can detect the 360-degree
movements of pedestrians, vehicles, road signs and traffic lights and create 3D maps. CV helps
them detect and analyse objects in real-time and take decisions like breaking, stopping or keep
driving.