Deep Learning Week 1 Assignment Answers Nptel

Last updated: January 16, 2026

Looking for NPTEL Deep Learning Week 1 Assignment Answers ? All week 1-12 Solution from NPTEL Assignment Solution Available

NPTEL Deep Learning Week 1 Assignment Answers (Jan-Apr 2026)

Que1. In the context of a signature descriptor, how is the “signature” of a shape generated?

a) By calculating the Fourier coefficients of the boundary points
b) By plotting the distance of boundary points from the centroid in various orientations
c) By recursively subdividing a shape into polygonal segments
d) None of the above

View Answer

Que2. Why is a logarithmic transformation used in the computation of Mel Frequency Cepstral Coefficients (MFCC) for audio signals?

a) To normalize the frequency spectrum
b) Because the human auditory system is more sensitive to signals that are not very loud
c) To remove high-frequency noise from the microphone
d) To convert the signal into a two-dimensional matrix

View Answer

Que3. Which of the following is a region descriptor?

a) Polygonal Representation
b) Fourier descriptor
c) Signature
d) Intensity histogram

View Answer

Que4. What is the primary difference between Traditional Machine Learning and Deep Learning regarding feature extraction?

a) Traditional machine learning ignores the raw signal
b) Deep learning requires the user to manually define and extract features before training
c) In deep learning, the machine learns to extract relevant features directly from the raw signal
d) Deep learning cannot process audio signals whereas traditional machine learning can

View Answer

Que5. In statistical moment calculation for a normalized histogram h(ri), what does the second-order moment (k = 2) represent?

a) The mean intensity
b) The skewness of the distribution
c) The maximum probability of the distribution
d) The variance (spread) of the histogram

View Answer

Que6. Signature descriptor of an unknown shape is given in the figure. Can you identify the unknown shape?

a) Circle
b) Square
c) Straight line
d) Rectangle

View Answer

Que7. Under what condition are the Bayes Minimum Risk Classifier and the Bayes Minimum Error Classifier mathematically identical?

a) When the loss function is a one-zero (or zero-one) loss function
b) When the prior probabilities of all classes are equal
c) When the feature space is exactly two-dimensional
d) When the likelihood ratio is close to 1

View Answer

Que8. Given an image I (Fig. 1), the gray co-occurrence matrix C (Fig. 2) can be constructed by specifying the displacement vector d = (dx, dy). Let the position operator be specified as (1,1), which has the interpretation: one pixel to the right and one pixel below. (Both the image and the partial gray co-occurrence matrix are given in the figures. Blank values and X in the gray co-occurrence matrix are unknown.) What is the value of X?

a) 0
b) 1
c) 2
d) 3

View Answer

Que9. A sub-segment of an arbitrary object boundary is represented by a discrete boundary function g(ri) after normalization. The distribution of boundary points relative to the centroid is characterized by the following discrete probability mass function:
Discrete radial distances: ri ∈ {2, 4, 6}
Corresponding probabilities: p(2) = 0.25, p(4) = 0.50, p(6) = 0.25

Calculate the second-order statistical moment of this boundary segment.

a) 2
b) 4
c) 4.5
d) 3

View Answer

Que10. When using Fourier Descriptors to reconstruct a square shape, what is the effect of using only low-order coefficients (P ≪ N)?

a) The reconstructed shape will have sharper corners
b) The reconstructed shape will lose detailed information (corners) and appear circular
c) The reconstruction will be identical to the original shape
d) The reconstruction will result in a single point

View Answer

NPTEL Deep Learning Week 1 Assignment Answers (Jan-Apr 2025)

Course Link: Click Here

Que. 1) Signature descriptor of an unknown shape is given in the figure. Can you identify the unknown shape?

a) Circle
b) Square
c) Straight line
d) Cannot be predicted

View Answer

Que. 2) To measure the smoothness, coarseness, and regularity of a region, we use which of the transformations to extract features?

a) Gabor Transformation
b) Wavelet Transformation
c) Both Gabor and Wavelet Transformation
d) None of the above

View Answer

Que. 3) Suppose the Fourier descriptor of a shape has KK coefficients, and we remove the last few coefficients, using only the first mm (m<Km < K) coefficients to reconstruct the shape. What will be the effect of using truncated Fourier descriptors on the reconstructed shape?

a) We will get a smoothed boundary version of the shape.
b) We will get only the fine details of the boundary of the shape.
c) The full shape will be reconstructed without any loss of information.
d) Low-frequency components of the boundary will be removed from the contour of the shape.

View Answer

Que. 4) While computing the polygonal descriptor of an arbitrary shape using the splitting technique, which of the following is taken as the starting guess?

a) Vertex joining the two closest points above a threshold on the boundary.
b) Vertex joining the two farthest points on the boundary.
c) Vertex joining any two arbitrary points on the boundary.
d) None of the above

View Answer

Que. 5) Consider a two-class Bayes’ Minimum Risk Classifier. The probabilities of classes W1W_1 and W2W_2 are P(W1)=0.3P(W_1) = 0.3 and P(W2)=0.7P(W_2) = 0.7, respectively. Given: P(x)=0.545P(x) = 0.545, P(x∣W1)=0.65P(x | W_1) = 0.65, and P(x∣W2)=0.25P(x | W_2) = 0.25. The loss matrix values are [2 1; 1 2]. If the classifier assigns xx to class W1W_1, then which of the following is true?

a) <1.79< 1.79
b) >1.79> 1.79
c) =1.09= 1.09
d) >1.09> 1.09

View Answer

Que. 6) The Fourier transformation of a complex sequence of numbers s(k)s(k) for k=0,…,N−1k = 0, …, N-1 is given by:

a) s(k)e−j2πk/Ns(k) e^{-j2\pi k/N}
b) (missing option)
c) (missing option)
d) (missing option)

View Answer

Que. 7) The gray co-occurrence matrix CC of an unknown image is given below. What is the value of the maximum probability descriptor?

Matrix CC:

1   2   2   2  
1   3   2   2  
2   2   2   2

a) 317\frac{3}{17}
b) 112\frac{1}{12}
c) 316\frac{3}{16}
d) 516\frac{5}{16}

View Answer

Que. 8) Which of the following is not a boundary descriptor?

a) Polygonal Representation
b) Fourier Descriptor
c) Signature
d) Histogram

View Answer

Que. 9) We use the gray co-occurrence matrix to extract which type of information?

a) Boundary
b) Texture
c) MFCC
d) Zero Crossing Rate

View Answer

Que. 10) If the larger values of the gray co-occurrence matrix are concentrated around the main diagonal, then which one of the following will be true?

a) The value of the element difference moment will be low.
b) The value of the inverse element difference moment will be low.
c) The value of entropy will be very low.
d) None of the above

View Answer