Artificial Intelligence

CBSE Class 10 AI Unit 4 MCQs with Answers & Explanations (2026–27) | Computer Vision

Class 10 · Artificial Intelligence

Preparing for the CBSE Class 10 Artificial Intelligence (Code 417) examination? This comprehensive collection of 50 Multiple Choice Questions (MCQs) from Unit 5 – Computer Vision is designed according to the latest CBSE syllabus for the 2026–27 session.

These MCQs cover all important concepts including the basics of Computer Vision, image processing, pixels, image resolution, RGB and grayscale images, Computer Vision tasks, Convolutional Neural Networks (CNNs), OpenCV, Orange Data Mining, Teachable Machine, and real-world AI applications. Every question includes the correct answer, a detailed explanation, and reasons why the remaining options are incorrect, helping students strengthen both conceptual understanding and examination skills.

Topics Covered

  • Introduction to Computer Vision
  • Computer Vision vs Image Processing
  • Applications of Computer Vision
  • Pixels, Resolution & Megapixels
  • Grayscale and RGB Images
  • Computer Vision Tasks (Classification, Localization, Detection & Segmentation)
  • Image Features
  • Convolution & Kernels
  • Convolutional Neural Networks (CNN)
  • ReLU, Pooling & Fully Connected Layers
  • OpenCV Library
  • Orange Data Mining
  • Teachable Machine & Lobe.ai
  • Case Study-Based Questions
Exam Tip: Computer Vision is one of the most important units in the CBSE AI syllabus. Focus on understanding the differences between Image Processing and Computer Vision, Computer Vision tasks, RGB image representation, CNN architecture, and popular AI tools like OpenCV and Teachable Machine. Competency-based and case-study questions are frequently asked from this unit.

Multiple Choice Questions (MCQs)

MCQs 1–10

Q1. What is the primary definition of Computer Vision (CV)?

A. A domain of AI that enables machines to speak human languages.
B. A domain of AI that enables machines to analyze visual information.
C. A tool for managing spreadsheets and statistical data.
D. A hardware component used only in DSLR cameras.

Show Answer & Explanation

Correct Answer: B. A domain of AI that enables machines to analyze visual information.

Explanation: Computer Vision is a branch of Artificial Intelligence that enables computers to acquire, process, and understand images and videos, allowing them to make decisions based on visual information.

Why other options are incorrect:

  • A. This describes Natural Language Processing (NLP), not Computer Vision.
  • C. Spreadsheet management is unrelated to Computer Vision.
  • D. A camera is only an input device, not Computer Vision itself.

Q2. Which of the following can be used as an input source for a Computer Vision system?

A. Infrared sensors
B. Text documents only
C. Binary code scripts
D. Audio recordings only

Show Answer & Explanation

Correct Answer: A. Infrared sensors

Explanation: Computer Vision systems can process images from cameras, videos, thermal cameras, and infrared sensors.

Why other options are incorrect:

  • B. Text documents are processed by NLP.
  • C. Binary scripts are not visual inputs.
  • D. Audio is handled by speech processing systems.

Q3. Computer Vision is considered a superset of:

A. Image Processing
B. Natural Language Processing
C. Robotics
D. Data Science

Show Answer & Explanation

Correct Answer: A. Image Processing

Explanation: Image Processing focuses on enhancing or modifying images, while Computer Vision goes further by understanding and interpreting them.

Why other options are incorrect:

  • B. NLP deals with text and language.
  • C. Robotics is a separate field that may use Computer Vision.
  • D. Data Science is a broader discipline.

Q4. What is the primary objective of Image Processing?

A. Enhancing raw images or preparing them for analysis
B. Predicting future events
C. Understanding human emotions
D. Translating text into speech

Show Answer & Explanation

Correct Answer: A. Enhancing raw images or preparing them for analysis

Explanation: Image Processing improves image quality through operations like resizing, filtering, brightness adjustment, and noise removal.

Why other options are incorrect:

  • B. Prediction belongs to machine learning tasks.
  • C. Emotion recognition is an application of AI, not image processing.
  • D. Translation is handled using NLP.

Q5. Which Computer Vision technology enables Google Translate to instantly translate text captured by a camera?

A. Linear Regression
B. Sentiment Analysis
C. Optical Character Recognition (OCR)
D. Neural Search

Show Answer & Explanation

Correct Answer: C. Optical Character Recognition (OCR)

Explanation: OCR detects and extracts text from images, allowing applications such as Google Translate to translate it in real time.

Why other options are incorrect:

  • A. Linear Regression predicts numerical values.
  • B. Sentiment Analysis determines opinions in text.
  • D. Neural Search is unrelated to text recognition.

Q6. Face Filters on social media platforms work by:

A. Detecting facial features and placing virtual objects
B. Increasing internet speed
C. Tracking browsing history
D. Measuring heart rate

Show Answer & Explanation

Correct Answer: A. Detecting facial features and placing virtual objects

Explanation: Face filters use Computer Vision to detect facial landmarks and accurately position digital effects.

Why other options are incorrect:

  • B. Internet speed is unrelated.
  • C. Browsing history is unrelated.
  • D. Face filters do not measure heart rate.

Q7. Resolution of a digital image is usually represented as:

A. Width × Height in pixels
B. Image size in MB
C. Number of colors only
D. Brightness level

Show Answer & Explanation

Correct Answer: A. Width × Height in pixels

Explanation: Resolution indicates the number of pixels along the width and height of an image, such as 1920 × 1080.

Why other options are incorrect:

  • B. File size depends on compression.
  • C. Resolution is independent of color count.
  • D. Brightness is unrelated to resolution.

Q8. The smallest unit of a digital image is called a:

A. Bit
B. Pixel
C. Node
D. Kernel

Show Answer & Explanation

Correct Answer: B. Pixel

Explanation: A pixel (picture element) is the smallest unit that forms a digital image.

Why other options are incorrect:

  • A. A bit stores binary data.
  • C. Nodes are used in networks and graphs.
  • D. Kernels are filters used in image processing.

Q9. In an 8-bit grayscale image, pixel values range from:

A. 0 to 100
B. 1 to 256
C. 0 to 255
D. –255 to +255

Show Answer & Explanation

Correct Answer: C. 0 to 255

Explanation: An 8-bit image contains 256 intensity levels ranging from 0 (black) to 255 (white).

Why other options are incorrect:

  • A. This is not the standard grayscale range.
  • B. The range starts from 0, not 1.
  • D. Negative values are not used.

Q10. In a grayscale image, a pixel value of 0 represents:

A. White
B. Gray
C. Transparent
D. Black

Show Answer & Explanation

Correct Answer: D. Black

Explanation: In grayscale images, 0 represents complete darkness (black), while 255 represents white.

Why other options are incorrect:

  • A. White corresponds to 255.
  • B. Gray lies between 0 and 255.
  • C. Transparency is not represented in grayscale values.

MCQs 11–20

Q11. One Megapixel (MP) is equal to:

A. 1,000 pixels
B. 10,000 pixels
C. 1,000,000 pixels
D. 100,000 pixels

Show Answer & Explanation

Correct Answer: C. 1,000,000 pixels

Explanation: One Megapixel represents one million pixels. Cameras with higher megapixels generally capture more image details.

Why other options are incorrect:

  • A. 1,000 pixels is too small.
  • B. 10,000 pixels is incorrect.
  • D. 100,000 pixels is not equal to one megapixel.

Q12. Which Computer Vision task assigns a single label to an entire image?

A. Object Detection
B. Image Classification
C. Instance Segmentation
D. Localization

Show Answer & Explanation

Correct Answer: B. Image Classification

Explanation: Image Classification predicts one category for the complete image, such as "Cat", "Dog", or "Car".

Why other options are incorrect:

  • A. Object Detection identifies multiple objects.
  • C. Instance Segmentation labels every pixel.
  • D. Localization identifies the position of an object.

Q13. Classification + Localization is mainly used when:

A. Multiple objects need to be detected.
B. Only one object is present and its location is required.
C. Every pixel must be labeled.
D. Images are converted to grayscale.

Show Answer & Explanation

Correct Answer: B. Only one object is present and its location is required.

Explanation: This task identifies what the object is and where it is located using a bounding box.

Why other options are incorrect:

  • A. Multiple objects require Object Detection.
  • C. Pixel labeling belongs to Instance Segmentation.
  • D. Grayscale conversion is image processing.

Q14. Which Computer Vision task identifies multiple objects and draws bounding boxes around them?

A. Image Classification
B. Object Detection
C. Image Enhancement
D. Image Compression

Show Answer & Explanation

Correct Answer: B. Object Detection

Explanation: Object Detection locates and classifies multiple objects present in an image or video.

Why other options are incorrect:

  • A. Classification labels the entire image.
  • C. Enhancement improves image quality.
  • D. Compression reduces file size.

Q15. Instance Segmentation is different because it:

A. Detects objects and labels every pixel belonging to them.
B. Only improves image brightness.
C. Removes image noise.
D. Reduces image size.

Show Answer & Explanation

Correct Answer: A. Detects objects and labels every pixel belonging to them.

Explanation: Instance Segmentation provides precise boundaries of every detected object instead of only drawing rectangles.

Why other options are incorrect:

  • B. Brightness adjustment is image processing.
  • C. Noise removal is image enhancement.
  • D. Image resizing is unrelated.

Q16. A grayscale image is stored internally as:

A. Three separate color channels
B. A two-dimensional array of pixel values
C. Audio signals
D. Binary text

Show Answer & Explanation

Correct Answer: B. A two-dimensional array of pixel values

Explanation: Grayscale images contain only intensity values arranged in rows and columns.

Why other options are incorrect:

  • A. Three channels are used in RGB images.
  • C. Audio signals are unrelated.
  • D. Images are not stored as text.

Q17. Which three primary colors are used to represent digital color images?

A. Red, Yellow, Blue
B. Cyan, Magenta, Yellow
C. Red, Green, Blue
D. Black, White, Gray

Show Answer & Explanation

Correct Answer: C. Red, Green, Blue

Explanation: RGB is the standard color model used in digital displays and images.

Why other options are incorrect:

  • A. These are primary colors for painting.
  • B. CMY is mainly used in printing.
  • D. These are grayscale colors.

Q18. In an RGB image, each pixel contains:

A. One value
B. Two values
C. Three values
D. Eight values

Show Answer & Explanation

Correct Answer: C. Three values

Explanation: Every RGB pixel stores the intensity values of Red, Green, and Blue channels.

Why other options are incorrect:

  • A. One value is used in grayscale images.
  • B. RGB requires three values.
  • D. Eight values are not stored for each pixel.

Q19. Which RGB combination produces White color?

A. (0, 0, 0)
B. (255, 255, 255)
C. (255, 0, 0)
D. (0, 255, 0)

Show Answer & Explanation

Correct Answer: B. (255, 255, 255)

Explanation: Full intensity of Red, Green, and Blue together produces White.

Why other options are incorrect:

  • A. Represents Black.
  • C. Represents Red.
  • D. Represents Green.

Q20. Which no-code Google tool allows users to build AI models using images, sounds, and poses?

A. OpenCV
B. Orange Data Mining
C. Teachable Machine
D. TensorFlow

Show Answer & Explanation

Correct Answer: C. Teachable Machine

Explanation: Google's Teachable Machine enables users to train image, audio, and pose recognition models without writing code.

Why other options are incorrect:

  • A. OpenCV is a programming library.
  • B. Orange is a visual data mining platform.
  • D. TensorFlow is a programming framework.

MCQs 21–30

Q21. Lobe.ai is best described as a:

A. Programming language for AI
B. No-code AutoML tool for image classification
C. Spreadsheet software
D. Professional image editing software

Show Answer & Explanation

Correct Answer: B. No-code AutoML tool for image classification

Explanation: Lobe.ai is a free, no-code platform developed by Microsoft that allows users to train image classification models without programming.

Why other options are incorrect:

  • A. Lobe.ai is not a programming language.
  • C. It is not spreadsheet software.
  • D. It is not designed for image editing.

Q22. In the Orange Data Mining case study, the AI model was developed to identify:

A. Different species of fish
B. Penguin weight
C. Bleached and healthy corals
D. Ocean temperature

Show Answer & Explanation

Correct Answer: C. Bleached and healthy corals

Explanation: The Coral Bleaching case study demonstrates how AI can classify coral images to help protect marine ecosystems.

Why other options are incorrect:

  • A. Fish classification is not discussed in this case study.
  • B. Penguin datasets belong to another unit.
  • D. The model does not predict water temperature.

Q23. In Computer Vision, an image feature refers to:

A. Camera hardware
B. Important information such as edges, corners, or blobs
C. File size of an image
D. Digital watermark

Show Answer & Explanation

Correct Answer: B. Important information such as edges, corners, or blobs

Explanation: Image features are distinctive parts of an image that help machines recognize and identify objects.

Why other options are incorrect:

  • A. Hardware captures images but is not a feature.
  • C. File size is unrelated to image features.
  • D. A watermark is not an image feature.

Q24. Which image feature is considered the best because it is unique and easy to locate?

A. Flat surfaces
B. Straight edges
C. Corners
D. Shadows

Show Answer & Explanation

Correct Answer: C. Corners

Explanation: Corners are highly distinctive because movement in any direction changes their appearance, making them easier to detect.

Why other options are incorrect:

  • A. Flat regions lack distinctive information.
  • B. Edges often appear similar along their length.
  • D. Shadows vary with lighting conditions.

Q25. Why are straight edges considered weaker features than corners?

A. They are always brighter.
B. They look similar throughout their length.
C. They exist only in grayscale images.
D. They contain no pixels.

Show Answer & Explanation

Correct Answer: B. They look similar throughout their length.

Explanation: Since edges appear similar along the line, locating an exact point is more difficult than locating a corner.

Why other options are incorrect:

  • A. Brightness has no relation to edge uniqueness.
  • C. Edges exist in all image types.
  • D. Every edge consists of pixels.

Q26. What is Convolution in Computer Vision?

A. A mathematical operation between an image and a kernel
B. A method for deleting images
C. A photo printing technique
D. A camera focusing mechanism

Show Answer & Explanation

Correct Answer: A. A mathematical operation between an image and a kernel

Explanation: Convolution applies a kernel to an image to detect features such as edges, patterns, and textures.

Why other options are incorrect:

  • B. Convolution does not delete images.
  • C. It is unrelated to printing.
  • D. Camera focusing is a hardware function.

Q27. A Kernel (or Filter) is:

A. The processor of a camera
B. A small matrix used during convolution
C. A memory card
D. A camera lens

Show Answer & Explanation

Correct Answer: B. A small matrix used during convolution

Explanation: Kernels slide across an image and perform mathematical operations to extract useful image features.

Why other options are incorrect:

  • A. The processor executes instructions.
  • C. Memory cards store data.
  • D. Lenses capture light only.

Q28. Which technique helps maintain the same output image size after convolution?

A. Cropping
B. Padding
C. Compression
D. Rotation

Show Answer & Explanation

Correct Answer: B. Padding

Explanation: Padding adds extra pixels around the image boundary so the output image retains the original dimensions.

Why other options are incorrect:

  • A. Cropping reduces image size.
  • C. Compression changes file size, not dimensions.
  • D. Rotation changes orientation only.

Q29. What does CNN stand for?

A. Central Neural Node
B. Computer Network Node
C. Convolutional Neural Network
D. Calculated Neural Number

Show Answer & Explanation

Correct Answer: C. Convolutional Neural Network

Explanation: CNN is a deep learning architecture specifically designed for image recognition and Computer Vision tasks.

Why other options are incorrect:

  • A. This is not a recognized AI term.
  • B. It relates to networking, not Computer Vision.
  • D. This expansion is incorrect.

Q30. Which layer is always the first layer in a Convolutional Neural Network?

A. Pooling Layer
B. Fully Connected Layer
C. Convolution Layer
D. Output Layer

Show Answer & Explanation

Correct Answer: C. Convolution Layer

Explanation: The Convolution Layer extracts important features such as edges, textures, and shapes from the input image.

Why other options are incorrect:

  • A. Pooling follows convolution.
  • B. Fully Connected Layer appears near the end.
  • D. The Output Layer gives the final prediction.

MCQs 31–40

Q31. The output generated by a Convolution Layer is called a:

A. Probability Table
B. Feature Map (Activation Map)
C. Kernel Vector
D. Pixel Matrix

Show Answer & Explanation

Correct Answer: B. Feature Map (Activation Map)

Explanation: After applying kernels during convolution, the extracted features are stored in a Feature Map, also known as an Activation Map.

Why other options are incorrect:

  • A. Probability tables are generated at the final output layer.
  • C. Kernel is the filter itself, not the output.
  • D. Pixel matrix refers to the original image data.

Q32. What is the primary function of the ReLU (Rectified Linear Unit) layer?

A. Increase image resolution
B. Replace negative values with zero
C. Convert RGB images to grayscale
D. Flatten the image

Show Answer & Explanation

Correct Answer: B. Replace negative values with zero

Explanation: The ReLU activation function introduces non-linearity by converting all negative values into zero while retaining positive values.

Why other options are incorrect:

  • A. ReLU does not change image resolution.
  • C. It does not perform color conversion.
  • D. Flattening is done before the Fully Connected Layer.

Q33. Why is the ReLU function applied in CNN?

A. To make important features more prominent
B. To reduce camera brightness
C. To increase image size
D. To compress image files

Show Answer & Explanation

Correct Answer: A. To make important features more prominent

Explanation: ReLU enhances useful image features and introduces non-linearity, enabling CNNs to learn complex patterns.

Why other options are incorrect:

  • B. Brightness adjustment is image processing.
  • C. ReLU does not change image dimensions.
  • D. Compression is unrelated.

Q34. What is the main purpose of the Pooling Layer?

A. Increase image resolution
B. Reduce the size of the Feature Map
C. Convert images into text
D. Add more pixels to an image

Show Answer & Explanation

Correct Answer: B. Reduce the size of the Feature Map

Explanation: Pooling reduces the spatial dimensions of Feature Maps, making the model faster and less prone to overfitting.

Why other options are incorrect:

  • A. Pooling does not improve resolution.
  • C. Text conversion is performed using OCR.
  • D. Pooling actually reduces information.

Q35. Which layer produces the final classification result in a CNN?

A. Convolution Layer
B. Pooling Layer
C. Fully Connected Layer
D. ReLU Layer

Show Answer & Explanation

Correct Answer: C. Fully Connected Layer

Explanation: The Fully Connected Layer uses extracted features to predict the final output class.

Why other options are incorrect:

  • A. Convolution extracts features.
  • B. Pooling reduces Feature Map size.
  • D. ReLU performs activation only.

Q36. Before entering the Fully Connected Layer, Feature Maps are:

A. Rotated
B. Flattened into a single vector
C. Converted into grayscale
D. Deleted

Show Answer & Explanation

Correct Answer: B. Flattened into a single vector

Explanation: Flattening converts the two-dimensional Feature Maps into a one-dimensional vector for classification.

Why other options are incorrect:

  • A. Rotation is unrelated.
  • C. Images are already processed before this stage.
  • D. Features are preserved, not deleted.

Q37. In facial recognition systems, Feature Maps help by:

A. Increasing camera resolution
B. Focusing on important facial features like eyes and nose
C. Saving images to cloud storage
D. Reducing image brightness

Show Answer & Explanation

Correct Answer: B. Focusing on important facial features like eyes and nose

Explanation: Feature Maps highlight important characteristics that help distinguish one face from another.

Why other options are incorrect:

  • A. Resolution is determined by the camera.
  • C. Feature Maps are not used for cloud storage.
  • D. Brightness adjustment is unrelated.

Q38. Which CNN layer makes the model less sensitive to small shifts in object position?

A. ReLU Layer
B. Input Layer
C. Pooling Layer
D. Kernel Layer

Show Answer & Explanation

Correct Answer: C. Pooling Layer

Explanation: Pooling creates a condensed representation of features, making CNNs more robust to small translations in images.

Why other options are incorrect:

  • A. ReLU only introduces non-linearity.
  • B. Input Layer receives the image.
  • D. Kernel is not a CNN layer.

Q39. TensorFlow and Keras are examples of:

A. Programming languages
B. Python libraries used for AI and Computer Vision
C. Image editing software
D. Database management systems

Show Answer & Explanation

Correct Answer: B. Python libraries used for AI and Computer Vision

Explanation: TensorFlow and Keras are popular deep learning libraries used to build and train CNN models.

Why other options are incorrect:

  • A. Python is the programming language; TensorFlow and Keras are libraries.
  • C. They are not image editing tools.
  • D. They are unrelated to database management.

Q40. OpenCV is primarily used for:

A. Database management
B. Real-time Computer Vision and Image Processing
C. Spreadsheet analysis
D. Financial forecasting

Show Answer & Explanation

Correct Answer: B. Real-time Computer Vision and Image Processing

Explanation: OpenCV is an open-source library widely used for image processing, object detection, face recognition, and real-time Computer Vision applications.

Why other options are incorrect:

  • A. OpenCV is not a DBMS.
  • C. Spreadsheet analysis is done using tools like Excel.
  • D. Financial forecasting is unrelated to OpenCV.

MCQs 41–50

Q41. Assertion (A): Corners are better image features than flat surfaces.
Reason (R): Corners are unique and can be accurately located, whereas flat regions appear similar over a large area.

A. Both A and R are true, and R is the correct explanation of A.
B. Both A and R are true, but R is not the correct explanation of A.
C. A is true, but R is false.
D. A is false, but R is true.

Show Answer & Explanation

Correct Answer: A. Both A and R are true, and R is the correct explanation of A.

Explanation: Corners are considered the best image features because they are unique and can be identified accurately, unlike flat regions that appear similar everywhere.

Why other options are incorrect:

  • B. The Reason correctly explains the Assertion.
  • C. The Reason is true.
  • D. The Assertion is also true.

Q42. What color is represented by the RGB value (0, 0, 0)?

A. White
B. Red
C. Black
D. Gray

Show Answer & Explanation

Correct Answer: C. Black

Explanation: When the intensity of Red, Green, and Blue is zero, no light is produced, resulting in the color black.

Why other options are incorrect:

  • A. White is represented by (255, 255, 255).
  • B. Red is represented by (255, 0, 0).
  • D. Gray contains equal non-zero RGB values.

Q43. In self-driving cars, Computer Vision is mainly used to:

A. Play music for passengers
B. Detect obstacles and recognize traffic signs
C. Predict fuel prices
D. Monitor internet speed

Show Answer & Explanation

Correct Answer: B. Detect obstacles and recognize traffic signs

Explanation: Computer Vision enables autonomous vehicles to understand road conditions, identify obstacles, recognize traffic signals, and make safe driving decisions.

Why other options are incorrect:

  • A. Music playback is unrelated to Computer Vision.
  • C. Fuel price prediction is not a CV application.
  • D. Internet speed monitoring has no connection with Computer Vision.

Q44. In the comparison between Human Vision and Computer Vision, the "Eye" corresponds to the:

A. CPU
B. Camera or Sensing Device
C. Hard Disk
D. RAM

Show Answer & Explanation

Correct Answer: B. Camera or Sensing Device

Explanation: Just as the human eye captures visual information, a camera or sensing device captures images for a Computer Vision system.

Why other options are incorrect:

  • A. The CPU processes data rather than capturing it.
  • C. A hard disk stores data.
  • D. RAM temporarily stores data during processing.

Q45. Which Orange Data Mining widget is used to visualize the relationship between two variables?

A. File Widget
B. Scatter Plot Widget
C. Test & Score Widget
D. Data Table Widget

Show Answer & Explanation

Correct Answer: B. Scatter Plot Widget

Explanation: The Scatter Plot widget helps users visualize relationships, trends, and patterns between two variables.

Why other options are incorrect:

  • A. File Widget imports datasets.
  • C. Test & Score evaluates machine learning models.
  • D. Data Table displays data records.

Q46. A computer stores a colored image as:

A. Text labels only
B. Numerical intensity values across Red, Green, and Blue channels
C. Audio signals
D. Binary sentences

Show Answer & Explanation

Correct Answer: B. Numerical intensity values across Red, Green, and Blue channels

Explanation: Every pixel in a colored image stores three numerical values representing the intensity of Red, Green, and Blue.

Why other options are incorrect:

  • A. Images are not stored as text.
  • C. Audio signals are unrelated.
  • D. Images are represented as numerical pixel values.

Q47. Another name for a Feature Map is:

A. Kernel
B. Activation Map
C. Pooling Matrix
D. Pixel Grid

Show Answer & Explanation

Correct Answer: B. Activation Map

Explanation: The terms Feature Map and Activation Map are used interchangeably in Convolutional Neural Networks.

Why other options are incorrect:

  • A. A kernel is the filter applied during convolution.
  • C. Pooling Matrix is not the output of convolution.
  • D. Pixel Grid refers to the original image.

Q48. An image has a resolution of 100 × 100 pixels. How many pixels does it contain?

A. 100
B. 1,000
C. 10,000
D. 100,000

Show Answer & Explanation

Correct Answer: C. 10,000

Explanation: Total number of pixels = Width × Height = 100 × 100 = 10,000 pixels.

Why other options are incorrect:

  • A. Incorrect calculation.
  • B. Incorrect multiplication.
  • D. Exceeds the actual number of pixels.

Q49. Which Computer Vision task is the most detailed because it assigns a label to every pixel of an object?

A. Image Classification
B. Object Detection
C. Instance Segmentation
D. Localization

Show Answer & Explanation

Correct Answer: C. Instance Segmentation

Explanation: Instance Segmentation identifies objects and classifies every pixel belonging to each object, providing highly accurate boundaries.

Why other options are incorrect:

  • A. Classification labels the entire image.
  • B. Object Detection uses bounding boxes only.
  • D. Localization identifies the position of a single object.

Q50. During the ReLU operation in a CNN, negative values are:

A. Converted to positive values
B. Multiplied by 255
C. Replaced with zero
D. Sent directly to the output layer

Show Answer & Explanation

Correct Answer: C. Replaced with zero

Explanation: The Rectified Linear Unit (ReLU) activation function removes negative values by replacing them with zero while keeping positive values unchanged.

Why other options are incorrect:

  • A. Negative values are not converted into positive values.
  • B. ReLU does not multiply values.
  • D. Negative values are eliminated before further processing.

? Quick Revision

  • Computer Vision enables machines to understand and interpret images and videos.
  • Image Processing improves images, while Computer Vision extracts meaningful information from them.
  • OCR (Optical Character Recognition) converts text present in images into editable text.
  • Image Classification assigns one label to an entire image.
  • Object Detection identifies multiple objects using bounding boxes.
  • Instance Segmentation labels every pixel belonging to each detected object.
  • RGB Images consist of Red, Green, and Blue color channels.
  • Convolution, ReLU, Pooling, and Fully Connected Layers are the major components of a CNN.
  • OpenCV, TensorFlow, Keras, Orange Data Mining, Teachable Machine, and Lobe.ai are popular Computer Vision tools.
  • Computer Vision is widely used in self-driving cars, facial recognition, medical imaging, OCR, surveillance, and augmented reality.

? Exam Tip

For the CBSE Class 10 Artificial Intelligence (Code 417) examination, focus on the differences between Computer Vision and Image Processing, understand the various Computer Vision tasks, revise RGB and grayscale images, learn the working of Convolutional Neural Networks (CNN), and remember the applications of tools like OpenCV, Teachable Machine, and Lobe.ai. Competency-based and case-study questions from these topics are frequently asked in the CBSE examination.