1. What is the primary goal of image classification in digital image processing?
(A) Image enhancement
(B) Object detection
(C) Assigning labels to image pixels or regions
(D) Image compression
2. Which of the following is a supervised learning technique used in image classification?
(A) K-Means
(B) PCA
(C) SVM
(D) Histogram Equalization
3. What does CNN stand for in image classification tasks?
(A) Computational Neural Network
(B) Central Neural Network
(C) Convolutional Neural Network
(D) Conditional Neural Network
4. Which layer in a CNN is primarily responsible for feature extraction?
(A) Fully connected layer
(B) Pooling layer
(C) Convolutional layer
(D) Output layer
5. What is the role of the pooling layer in a convolutional neural network?
(A) Increases resolution
(B) Reduces spatial size
(C) Adds more filters
(D) Normalizes data
6. Which pooling method is most commonly used in CNNs?
(A) Max pooling
(B) Min pooling
(C) Mean pooling
(D) Sum pooling
7. Which activation function is commonly used in deep learning for image classification?
(A) Sigmoid
(B) ReLU
(C) Tanh
(D) Softmax
8. What is the purpose of the softmax layer in a classification network?
(A) Reduce dimensions
(B) Apply convolution
(C) Perform normalization
(D) Output class probabilities
9. Which term refers to the error between predicted and actual labels in classification?
(A) Entropy
(B) Bias
(C) Loss
(D) Variance
10. Which loss function is commonly used in multi-class classification tasks?
(A) Mean Squared Error
(B) Binary Cross Entropy
(C) Categorical Cross Entropy
(D) Hinge Loss
11. Which method splits the dataset into training and testing sets?
(A) Histogram Matching
(B) Data Augmentation
(C) Cross-validation
(D) Image Segmentation
12. Which of the following techniques improves generalization in CNNs?
(A) Overfitting
(B) Dropout
(C) Max pooling
(D) Padding
13. Which metric is not typically used to evaluate classification performance?
(A) Precision
(B) Recall
(C) F1-Score
(D) PSNR
14. Which approach increases dataset size by transformations?
(A) Normalization
(B) Augmentation
(C) Pooling
(D) Classification
15. Which classifier is commonly used for binary image classification problems?
(A) K-Means
(B) SVM
(C) PCA
(D) Canny
16. What does the term “overfitting” mean in image classification?
(A) Model performs well on new data
(B) Model learns noise in training data
(C) Model underestimates training labels
(D) Model ignores input features
17. Which layer type flattens the output in CNNs for classification?
(A) Convolutional layer
(B) Pooling layer
(C) Fully connected layer
(D) Flatten layer
18. Which algorithm uses labeled data for classification tasks?
(A) K-Means
(B) DBSCAN
(C) KNN
(D) ICA
19. Which term defines the number of filters in a convolutional layer?
(A) Padding
(B) Kernel size
(C) Stride
(D) Depth
20. Which technique is not a dimensionality reduction method used in classification?
(A) PCA
(B) LDA
(C) t-SNE
(D) SVM
21. What is the role of the ReLU function in CNNs?
(A) Normalize data
(B) Reduce dimensions
(C) Introduce non-linearity
(D) Reduce noise
22. Which CNN architecture is known for its deep structure with 19 layers?
(A) AlexNet
(B) LeNet
(C) VGG-19
(D) ResNet
23. Which network introduced the concept of residual learning?
(A) AlexNet
(B) ResNet
(C) VGG
(D) GoogLeNet
24. What is transfer learning in the context of image classification?
(A) Learning from test data
(B) Transferring data across classes
(C) Using pre-trained models
(D) Switching models dynamically
25. Which layer is typically used at the end of a CNN for classification?
(A) Pooling
(B) Softmax
(C) ReLU
(D) Convolution
26. What does an epoch refer to in training a classification model?
(A) Single batch of data
(B) All model parameters
(C) One full pass of training data
(D) A single neuron activation
27. Which evaluation metric is best when dealing with imbalanced datasets?
(A) Accuracy
(B) Recall
(C) Precision
(D) F1-Score
28. What is the main function of a classification layer in CNN?
(A) Feature selection
(B) Noise reduction
(C) Assign class labels
(D) Normalization
29. Which CNN model won the ImageNet 2012 competition?
(A) VGGNet
(B) ResNet
(C) AlexNet
(D) Inception
30. Which type of learning is image classification based on?
(A) Unsupervised
(B) Semi-supervised
(C) Supervised
(D) Reinforcement
31. What is the main drawback of using high learning rates?
(A) Slow convergence
(B) Model underfitting
(C) Poor generalization
(D) Overshooting minima
32. Which of the following is not a component of CNN?
(A) Convolutional layer
(B) Pooling layer
(C) Decision tree
(D) Fully connected layer
33. Which of the following is a color image classification dataset?
(A) MNIST
(B) CIFAR-10
(C) COCO
(D) Pascal VOC
34. Which image classification task has multiple labels per image?
(A) Binary classification
(B) Multi-class classification
(C) Multi-label classification
(D) One-vs-all classification
35. Which of the following networks is best for real-time image classification?
(A) ResNet-152
(B) MobileNet
(C) VGG-19
(D) DenseNet
36. Which CNN model uses depthwise separable convolutions?
(A) LeNet
(B) MobileNet
(C) AlexNet
(D) VGGNet
37. Which evaluation metric measures the ratio of true positives to total predicted positives?
(A) Recall
(B) Accuracy
(C) Precision
(D) F1-Score
38. Which of the following is not a data preprocessing step?
(A) Resizing
(B) Normalization
(C) Label encoding
(D) Pooling
39. Which approach is used to prevent overfitting?
(A) Increasing training data
(B) Using deeper networks
(C) Reducing learning rate
(D) Ignoring dropout
40. What does the stride parameter in convolution determine?
(A) Number of filters
(B) Kernel size
(C) Movement of filter across input
(D) Padding size
41. Which dataset contains grayscale images of handwritten digits?
(A) CIFAR-10
(B) ImageNet
(C) MNIST
(D) COCO
42. Which type of classification involves only two possible classes?
(A) Multi-class
(B) Multi-label
(C) Binary
(D) Hierarchical
43. Which algorithm is best suited for non-linear image classification?
(A) Linear Regression
(B) Decision Tree
(C) Logistic Regression
(D) SVM with RBF kernel
44. What is the major difference between CNN and traditional neural networks for images?
(A) Use of pooling
(B) Use of backpropagation
(C) Fixed weights
(D) Fully connected layers only
45. What is the output of a softmax function?
(A) Binary values
(B) Normalized probabilities
(C) Raw logits
(D) Feature maps
46. Which of the following models is designed for very deep networks?
(A) VGGNet
(B) LeNet
(C) ResNet
(D) AlexNet
47. Which dataset is largest among the following for classification?
(A) MNIST
(B) CIFAR-10
(C) ImageNet
(D) Fashion-MNIST
48. Which CNN architecture introduced inception modules?
(A) AlexNet
(B) ResNet
(C) GoogLeNet
(D) VGGNet
49. Which technique modifies model weights based on error feedback?
(A) Pooling
(B) Convolution
(C) Backpropagation
(D) Padding
50. Which of the following is used to reduce the internal covariate shift during training in image classification?
(A) Dropout
(B) Batch Normalization
(C) Pooling
(D) Padding
