Image recognition projects quiz

Question 1

What is image recognition in simple terms?

Accepted Answer

A computer naming what is in a picture by learning from examples

Answer

A program that only stores photos in folders

Answer

A tool that makes pictures brighter

Answer

A way to print images faster

Question 2

How does a computer 'see' an image?

Accepted Answer

As a grid of pixels made of numbers

Answer

As a single word

Answer

As sound waves

Answer

As a smell

Question 3

What do RGB values represent in a digital image?

Accepted Answer

Red, Green, and Blue color amounts in a pixel

Answer

Round, Grey, Black shapes

Answer

Random Generated Bits

Answer

Resize, Group, Blur tools

Question 4

Why do image recognition projects need lots of example images?

Accepted Answer

So it can learn general patterns and work on new images

Answer

So the files take more disk space

Answer

Because pictures are decoration

Answer

To slow the computer down on purpose

Question 5

What is a label (or class) in an image recognition dataset?

Accepted Answer

The correct name attached to an image

Answer

The brightness of the image

Answer

The size of the file

Answer

The name of the camera

Question 6

What is the difference between training data and test data?

Accepted Answer

Training data is for learning; test data checks unseen performance

Answer

They are exactly the same thing

Answer

Test data is used first, training never

Answer

Training data is only for color photos

Question 7

What does 'accuracy' mean when judging an image classifier?

Accepted Answer

The fraction of images labeled correctly

Answer

The number of pixels in an image

Answer

How fast the model runs

Answer

The size of the dataset

Question 8

What is a neural network, described simply?

Accepted Answer

Connected units with weights that learn from examples

Answer

A network cable for computers

Answer

A drawing app

Answer

A spreadsheet of photos

Question 9

What kind of neural network is most popular for images, and why?

Accepted Answer

CNNs, because they scan patches to find image features

Answer

Email networks, because they are fast

Answer

Phone networks, because they are wireless

Answer

Spreadsheets, because they store numbers

Question 10

What is a 'feature' in image recognition?

Accepted Answer

A useful pattern like an edge or texture the model detects

Answer

The price of the image

Answer

The author of the photo

Answer

A type of camera lens

Question 11

What does it mean to 'train' an image recognition model?

Accepted Answer

Adjusting the model on labeled images to reduce mistakes

Answer

Deleting all the images

Answer

Printing the images

Answer

Renaming the image files

Question 12

Why is it important to resize all images to the same size before training?

Accepted Answer

Models expect a fixed, consistent input size

Answer

Bigger files always look nicer

Answer

It changes the image colors to grey

Answer

It deletes unwanted images

Question 13

What is data augmentation in image projects?

Accepted Answer

Making new training images by transforming existing ones

Answer

Deleting half the dataset

Answer

Turning images into text

Answer

Buying more cameras

Question 14

What is overfitting in an image recognition model?

Accepted Answer

The model memorizes training images and fails on new ones

Answer

The model is too small to open

Answer

The image files are too large

Answer

The screen shows the wrong colors

Question 15

What is the goal of splitting data into training, validation, and test sets?

Accepted Answer

To teach, tune, and fairly test the model on separate data

Answer

To make three copies of the same photos

Answer

To slow training down

Answer

To store images in three folders for fun

Question 16

What does 'classification' mean in image recognition?

Accepted Answer

Sorting an image into one category from a list

Answer

Measuring image file size

Answer

Changing image colors

Answer

Counting megabytes

Question 17

How is object detection different from image classification?

Accepted Answer

Detection locates objects with boxes; classification gives one label

Answer

They are identical

Answer

Detection only works on text

Answer

Classification draws boxes, detection does not

Question 18

What is a confidence score (probability) in a prediction?

Accepted Answer

A number showing how sure the model is about its guess

Answer

The number of pixels

Answer

The file format

Answer

The camera brand

Question 19

Why might converting an image to grayscale help some projects?

Accepted Answer

It simplifies data when color is not needed

Answer

It always improves color accuracy

Answer

It adds more colors

Answer

It makes files unreadable

Question 20

What is transfer learning and why is it useful for beginners?

Accepted Answer

Reusing a pretrained model to learn a new task with less data

Answer

Moving images between two computers

Answer

Copying files to a USB drive

Answer

Translating images into another language

Question 21

What is the MNIST dataset commonly used for in beginner projects?

Accepted Answer

Recognizing handwritten digits 0 to 9

Answer

Identifying songs

Answer

Detecting spam emails

Answer

Translating languages

Question 22

What is a 'class imbalance' problem in image datasets?

Accepted Answer

Some classes have many more images than others

Answer

All images are the same size

Answer

The screen is too bright

Answer

There are no labels at all

Question 23

Why should you be careful about privacy when collecting images of people?

Accepted Answer

Photos of people are personal and need consent

Answer

Photos never contain personal information

Answer

Privacy only matters for text

Answer

People enjoy having photos shared without asking

Question 24

What is bias in an image recognition model?

Accepted Answer

The model works unfairly better for some groups than others

Answer

The model runs too slowly

Answer

The images are too colorful

Answer

The dataset has no labels

Question 25

What is a real-world example of image recognition being helpful?

Accepted Answer

Helping doctors spot disease in medical scans

Answer

Making the internet slower

Answer

Storing recipes

Answer

Charging a phone battery

Question 26

In Python, which popular library is often used to build neural networks for images?

Accepted Answer

TensorFlow/Keras or PyTorch

Answer

Microsoft Word

Answer

A web browser

Answer

A music player

Question 27

What does normalizing pixel values (e.g., dividing by 255) do?

Accepted Answer

Rescales pixels to 0-1 to help the model learn better

Answer

Deletes dark pixels

Answer

Makes the image black and white

Answer

Adds more pixels

Question 28

What is a confusion matrix used for?

Accepted Answer

A table showing which classes the model confuses

Answer

A list of file names

Answer

A color palette

Answer

A type of camera setting

Question 29

Why can lighting and background affect an image recognition project?

Accepted Answer

They change pixel values and can confuse the model

Answer

They never affect digital images

Answer

They only matter for sound

Answer

They change the file format

Question 30

What is an 'epoch' during model training?

Accepted Answer

One full pass through the whole training dataset

Answer

A single pixel

Answer

One image file

Answer

The name of a model

Test yourself