Download 04-BIO-B4 Digital Image Processing Exam A and more Exams Technology in PDF only on Docsity!
Question 1. What is the primary objective of digital image processing? A) To create artistic images for visual appeal B) To perform operations on images to enhance, restore, or encode information C) To develop hardware for image acquisition D) To replace the need for human visual perception Answer: B Explanation: The main goal of digital image processing is to perform various operations on images to improve their quality, extract information, or encode data for storage and transmission. Question 2. Which of the following fields is most directly related to digital image processing? A) Quantum physics B) Computer vision C) Organic chemistry D) Mechanical engineering Answer: B Explanation: Computer vision is directly related because it involves interpreting and understanding images and video data, often using digital image processing techniques. Question 3. Which device primarily converts physical scenes into digital images? A) Microcontroller B) Sensor (e.g., camera or scanner) C) CPU processor D) Display monitor Answer: B Explanation: Sensors such as cameras and scanners capture physical scenes and convert them into digital image data. Question 4. The electromagnetic spectrum relevant to imaging primarily includes which range? A) Radio waves, microwaves, infrared, visible light, ultraviolet, X-rays, gamma rays B) Only visible light
C) Only infrared and ultraviolet D) Only X-rays and gamma rays Answer: A Explanation: Imaging utilizes a broad range of electromagnetic spectrum, with visible light being just a small part; other regions like infrared and ultraviolet are also relevant. Question 5. Sampling in digital imaging refers to: A) Converting an analog signal into a digital one by measuring its value at discrete points B) Changing the color depth of an image C) Reducing the image size by compression D) Increasing the resolution of an image using interpolation Answer: A Explanation: Sampling involves taking measurements of a continuous signal (or image) at discrete points to convert it into digital form. Question 6. Quantization in digital images pertains to: A) Assigning discrete intensity levels to sampled values B) Dividing the image into segments C) Increasing the spatial resolution of the image D) Smoothing the image to reduce noise Answer: A Explanation: Quantization is the process of mapping a range of sample values to a finite set of discrete levels, affecting the image's intensity resolution. Question 7. A pixel in a digital image is best described as: A) A tiny light source B) The smallest addressable element in an image with a specific intensity or color C) A type of sensor used in cameras D) The entire image itself Answer: B
Question 11. A gray-level transformation that produces a negative image is called: A) Log transformation B) Power-law (gamma) transformation C) Image negative transformation D) Piecewise linear transformation Answer: C Explanation: Negatives invert the intensity levels, producing an image where dark areas become light and vice versa. Question 12. Histogram equalization primarily aims to: A) Reduce image size B) Enhance contrast by redistributing intensity values C) Convert a grayscale image to color D) Smooth the image to reduce noise Answer: B Explanation: Histogram equalization enhances image contrast by spreading out the most frequent intensity values across the entire range. Question 13. Which filter is typically used for image smoothing? A) Laplacian filter B) Gaussian filter C) High-pass filter D) Unsharp mask Answer: B Explanation: Gaussian filters are low-pass filters used for smoothing images to reduce noise and details. Question 14. The primary purpose of a high-pass filter in spatial domain filtering is to: A) Blur the image B) Enhance edges and fine details
C) Reduce image noise D) Perform histogram equalization Answer: B Explanation: High-pass filters emphasize high-frequency components like edges, sharpening the image. Question 15. Which noise type is characterized by sparsely distributed bright and dark pixels? A) Gaussian noise B) Salt-and-Pepper noise C) Speckle noise D) Uniform noise Answer: B Explanation: Salt-and-Pepper noise manifests as random black and white pixels due to impulse disturbances. Question 16. The median filter is particularly effective against which type of noise? A) Gaussian noise B) Salt-and-Pepper noise C) Speckle noise D) Uniform noise Answer: B Explanation: Median filtering is effective in removing impulse noise like Salt-and-Pepper noise by replacing a pixel with the median of its neighborhood. Question 17. The Fourier Transform of an image converts it from the spatial domain into the: A) Color domain B) Frequency domain C) Time domain D) Spatial domain Answer: B
Question 21. The degradation model in image restoration includes which of the following? A) The original image only B) The degradation function and additive noise C) Only the noise component D) The compression algorithm used Answer: B Explanation: The degradation model considers how the original image is affected by a degradation function and additive noise during image capture or transmission. Question 22. Impulse noise is also known as: A) Gaussian noise B) Salt-and-Pepper noise C) Speckle noise D) Rayleigh noise Answer: B Explanation: Impulse noise appears as randomly distributed black and white pixels, characteristic of Salt- and-Pepper noise. Question 23. Which spatial filter is most effective for removing salt-and-pepper noise? A) Averaging filter B) Median filter C) Gaussian filter D) Laplacian filter Answer: B Explanation: The median filter effectively removes impulse noise by replacing each pixel with the median of its neighborhood. Question 24. Lossless compression techniques guarantee: A) Perfect reconstruction of the original image B) Slight loss of image quality for higher compression ratios
C) Reduction in image resolution D) Faster processing but lower quality Answer: A Explanation: Lossless compression ensures the original image can be perfectly reconstructed without any loss of information. Question 25. Huffman coding is based on: A) Fixed-length codes for all symbols B) Variable-length codes based on symbol probabilities C) Run-length encoding principles D) Transform coding techniques Answer: B Explanation: Huffman coding assigns shorter codes to more frequent symbols, optimizing compression based on symbol probability distributions. Question 26. Run-length encoding (RLE) is most effective for images with: A) High detail and noise B) Large areas of uniform intensity C) Random pixel intensities D) Complex textures Answer: B Explanation: RLE compresses sequences of identical pixels efficiently, making it suitable for images with large uniform regions. Question 27. In lossy compression, the transform coding step often uses: A) Fourier Transform B) Discrete Cosine Transform (DCT) C) Run-length encoding D) Huffman coding Answer: B
Question 31. Morphological dilation has the effect of: A) Shrinking objects in the image B) Filling small holes and connecting gaps in objects C) Removing small objects D) Smoothing object contours Answer: B Explanation: Dilation adds pixels to object boundaries, filling small holes and bridging gaps. Question 32. Morphological erosion is used to: A) Expand objects in the image B) Remove small objects and thin structures C) Fill holes within objects D) Connect disjoint objects Answer: B Explanation: Erosion shrinks objects by removing boundary pixels, useful for eliminating small noise or thinning structures. Question 33. The hit-or-miss transformation is primarily used for: A) Detecting specific patterns or shapes in an image B) Smoothing noisy images C) Segmenting regions based on intensity thresholds D) Enhancing edges Answer: A Explanation: It is a morphological operation used for pattern matching and shape detection. Question 34. Image segmentation is the process of: A) Compressing an image for storage B) Partitioning an image into meaningful regions or objects C) Converting a color image into grayscale
D) Enhancing image contrast Answer: B Explanation: Segmentation divides an image into regions or objects to analyze or interpret the image content. Question 35. Edge detection algorithms like Sobel and Prewitt are based on: A) Color histograms B) Gradient operators C) Fourier transforms D) Thresholding techniques Answer: B Explanation: They compute the gradient magnitude to identify regions with rapid intensity changes— edges. Question 36. The Canny edge detector involves which of the following steps? A) Smoothing, gradient calculation, non-maximum suppression, and hysteresis thresholding B) Thresholding, dilation, and erosion C) Fourier transform and filtering D) Histogram equalization and color segmentation Answer: A Explanation: Canny involves multiple steps including smoothing, gradient computation, suppression of non-maximum pixels, and hysteresis thresholding for robust edge detection. Question 37. Multi-level thresholding in image segmentation is used when: A) The image has more than two classes or regions B) The image contains only binary objects C) The image is in color D) The image is noisy Answer: A Explanation: Multi-level thresholding segments images into multiple regions based on intensity levels.
B) Equalize the histogram of an image C) Compress an image based on its histogram D) Remove noise from an image Answer: A Explanation: Histogram matching transforms the intensity distribution of an image to match a specified histogram, useful for standardizing images. Question 42. Which of the following is a lossless image compression technique? A) JPEG B) Huffman coding C) JPEG D) Discrete Cosine Transform (DCT) Answer: B Explanation: Huffman coding is a lossless compression method, while JPEG involves lossy compression. Question 43. The main disadvantage of lossy compression is: A) Loss of some original image information and possible quality degradation B) Increased file size C) Cannot be used for color images D) Limited to binary images only Answer: A Explanation: Lossy compression sacrifices some data to achieve higher compression ratios, potentially affecting image quality. Question 44. The Lempel-Ziv-Welch (LZW) algorithm is primarily used for: A) Lossless data compression in formats like GIF and TIFF B) Transform coding in JPEG C) Noise reduction in images D) Color space conversion
Answer: A Explanation: LZW is a lossless compression algorithm used in formats like GIF and TIFF. Question 45. The main advantage of transform coding in lossy compression is: A) It allows significant data reduction by discarding less perceptible frequency components B) It guarantees perfect reconstruction of the original image C) It is computationally faster than spatial domain techniques D) It works only on binary images Answer: A Explanation: Transform coding reduces data by removing frequency components less noticeable to human vision, enabling higher compression. Question 46. The RGB color model is based on: A) Additive mixing of light from three primary colors B) Subtractive mixing of inks or dyes C) Human perception of cone cells only D) The physical properties of the paper Answer: A Explanation: RGB is an additive model where colors are created by combining red, green, and blue light. Question 47. The main limitation of the RGB color model in image processing is: A) It does not align with human perception of color differences B) It cannot represent a wide range of colors C) It is only suitable for black-and-white images D) It is a subtractive color model Answer: A Explanation: RGB does not correspond directly to how humans perceive color differences, which can complicate certain processing tasks.
B) A filter used in Fourier domain filtering C) An image compression technique D) A type of color model Answer: A Explanation: A structuring element defines the neighborhood used in dilation, erosion, and related morphological operations. Question 52. Thinning and thickening are morphological algorithms used to: A) Reduce objects to skeletal forms and expand objects, respectively B) Enhance color saturation C) Perform histogram equalization D) Segment images based on intensity thresholds Answer: A Explanation: Thinning reduces objects to their skeletons, while thickening expands object boundaries, useful in shape analysis. Question 53. The importance of image segmentation in medical imaging is primarily to: A) Detect and isolate regions of interest such as tumors or organs B) Compress medical images for storage C) Convert images to binary form for printing D) Enhance color contrast for aesthetic purposes Answer: A Explanation: Segmentation helps in identifying and analyzing specific structures or abnormalities within medical images. Question 54. The gradient operators in edge detection include: A) Roberts, Prewitt, Sobel B) Fourier, Laplacian, Canny C) Histogram equalization, RLE, Huffman D) DCT, DFT, FFT
Answer: A Explanation: Roberts, Prewitt, and Sobel are common gradient-based operators for detecting edges based on intensity gradients. Question 55. The Laplacian of Gaussian (LoG) operator combines: A) Second derivative (Laplacian) and Gaussian smoothing B) First derivative and histogram equalization C) Fourier transform and thresholding D) Color space conversion and morphological filtering Answer: A Explanation: LoG smooths the image with a Gaussian filter and then applies the Laplacian for edge detection, reducing noise sensitivity. Question 56. Thresholding techniques are primarily used to: A) Convert grayscale images into binary images based on intensity criteria B) Detect edges in an image C) Compress images losslessly D) Remove noise from images Answer: A Explanation: Thresholding segments an image into two or more regions by classifying pixels based on intensity thresholds. Question 57. Adaptive thresholding differs from global thresholding because it: A) Calculates thresholds locally for different regions of the image B) Uses a fixed threshold value for the entire image C) Is only applicable to color images D) Does not depend on pixel intensity values Answer: A Explanation: Adaptive thresholding computes thresholds for local regions, better handling varying lighting conditions.
B) Store image spatial data C) Represent color information only D) Provide a compressed version of the image Answer: A Explanation: An image histogram depicts how pixel intensities are distributed, useful for analysis and enhancement. Question 62. Which of the following is a property of the Fourier Transform? A) Linearity B) Non-invertibility C) Time invariance only D) Spatial domain operation only Answer: A Explanation: The Fourier Transform is linear, meaning the transform of a sum of functions equals the sum of their transforms. Question 63. The main difference between spatial domain and frequency domain filtering is: A) Spatial filtering manipulates pixel values directly; frequency filtering modifies frequency components B) Spatial filtering is faster and more efficient than frequency filtering C) They are essentially the same processes D) Frequency domain filtering cannot be used for noise reduction Answer: A Explanation: Spatial domain filtering operates directly on pixel intensities, while frequency domain filtering modifies the image’s frequency spectrum. Question 64. The ideal low-pass filter in frequency domain: A) Passes all frequencies below a cutoff and blocks higher frequencies B) Passes all frequencies equally C) Passes only high frequencies D) Completely blocks all frequencies
Answer: A Explanation: An ideal low-pass filter allows low frequencies to pass, smoothing the image, and blocks high frequencies. Question 65. In image restoration, the degradation function H represents: A) The effect of the image formation process or distortion B) The original undistorted image C) The noise component only D) The compression algorithm used Answer: A Explanation: H models how the original image is degraded during acquisition or transmission, affecting the observed image. Question 66. The main limitation of inverse filtering in image restoration is: A) It amplifies noise, especially when H has small or zero values B) It is computationally very slow C) It cannot be used with noisy images D) It always produces perfect restoration regardless of noise Answer: A Explanation: Inverse filtering can significantly amplify noise when the degradation function has small magnitude values, leading to instability. Question 67. Gaussian noise is characterized by: A) Random variations with a bell-shaped probability distribution B) Sparse black and white pixels C) Multiplicative noise that depends on the signal intensity D) Uniform distribution across all pixels Answer: A Explanation: Gaussian noise follows a normal distribution, causing subtle, random variations in pixel intensities.
