Cyber Forensics — DU B.Sc. (Hons) Computer Science Previous Year Question Papers

Information and Image Retrieval 4. Parts of a question should be attempted together. QI (a) Name any two low-level features in lnformationRetrieval. 2 ap) Given the sentence: "Information retrieval systems retrieve relevant information 2 from documents." Identify and remove the stop words. Write the fmal token list. (c) Explain the Ranked Retrieval Model in Information Retrieval. 2 (d) Describe color moments jn CBIR systems and their role in image representation. 3 (c) Define invariant features. Why are they important in CBIR systems? 3 © Disouss any two dimensionality reduction techniques in improving retrieval 3 (g) Differentiate between stemming and lemmatization. (h) Given the following feature vectors: Discuss relevance feedback and its impact on retrieval performance. Explain any two texture features and shape features used in CBIR. D i : "data mining and data analysis" D2: "data science and machine leaning" D3 : "analysits of data systems" i. ii. Compute the Term Frequency (TF) of the term `.data" in D1. Compute the Inverse Document Frequency (mF) of the term "data". iii. Calculate the TF-IDF value of``data" in D1. `` q]) A retrieval system refums 30 documents, out of which 18 are relevant. The total number of documents in the collection is 45. i. Calculate precision and Recall €i. Compute the F-measure (a = 1) Q3 (a) (b) What is Average precision? For a query, the relevant documents are retrieved at ranks: 1, 3, and 6. Compute the Average Precision. Given the following grayscale image: i. Construct the Gray-Level co-occurrence Matrix (GLCM) for G = 90°and distance d = 1. ii. Normalize the GLCM. iii. Compute the Energy of the resultant matrix. Q4 (a) Explain the sARmodel in image retrieval. How does ithelp inrepresenting spatial relationships among image regions? (b) Explain the following similarity measures used in image retrieval systems.: 9 i. Modified Retrieval Rank (MRR) ii. Average Modified Retrieval Rank (AVMRR) Q5 (a) Explain the architecture of the QBIC (Query By Image Content) system in Content- 5 Based Image Retrieval systems. (b) Discuss the advantages and disadvantages of Text-Based Image Retrieval (TBIR) 5 (c) What is WebSEEK? Discuss its working and explain the limitations of WebSEEK in 5 Content-Based Image Retrieval systems. State curse of dimensionality. What is the necessity of dimensionality reduction. 4 Explain the autocorrelation in image texture analysis. 5 (b) An image has 2 colors: Red (R) and Blue (B). 6 (c) At distance d = 1, the following probabilities are given: Probability that a pixel of color R has a neighboring pixel of color R = 0.4 Probability that a pixel of color R has a neighboring pixel of color B = 0.6 Probability that a pixel of color B has a neighboring pixel of color R = 0.5 Probability that a pixel of color B has a neighboring pixel of color B = 0.5 i. Construct the color correlogram matrix. il. Identify the color pair that has the highest spatial correlation.