Content Based Image and Video Retrieval Algorithm
Download as PPTX, PDF3 likes2,705 views
The document describes content-based image and video retrieval (CBIR) algorithms. It discusses how CBIR works by extracting features from query images, indexing images, and retrieving similar images based on color, shape, and texture features. CBIR techniques include reverse image search, semantic retrieval using queries, and relevance feedback to refine searches based on user input about retrieved images. The document provides examples of CBIR applications in areas like crime prevention, military, web searching, and medical diagnosis.
Content Based Image and Video Retrieval Algorithm
- 1. 1 Content-Based Image and Video Retrieval Algorithm Submitted By: Akshit Kumar Bum B.E. in Information Technology (Final Year).
- 2. 2 CBIR is the application of computer vision technique to the image retrieval problem. The search uses the content of the image. Content might be colors, shapes, textures, or any other information derived from image itself. INTRODUCTION TO CBIR
- 3. 3 There is an amazing growth in the amount of digital image and video data in recent years. So the efficient image retrieval required. The limitations inherent in metadata-based systems. Old concept based approaches are time consuming and ineffective. NEED OF CBIR So much of irrelevant images Search in a different way So, the CBIR is Solution Seems Interesting, But How It Works ?????
- 4. HOW CBIR WORKS 4 Query Image Feature Extraction Initial Feedback Retrieved Similar Images
- 5. 5 Query Techniques Different implementation of CBIR make use of different user query. i.e. Reverse image search Semantic Retrieval Relevance Feedback(Human Interaction ) Iterative/Machine Learning Feature Extraction and indexing Extract the feature from queried image . i.e. Based on color Based on Shape Based on Texture Image indexing CBIR TECHNIQUES
- 6. 6 Provide the sample image to the CBIR system, and CBIR system will retrieve relevant images. QUERY TECHNIQUES Reverse image search CBIR SYSTEM
- 7. Query Techniques Reverse image search Uses Locate the Source of an image. Find higher resolution versions. Discover webpages where the image appears. Track down the content creator. Get information about an image. 7
- 8. 8 Google images TinEye images Application in popular search system
- 9. 9 Semantic retrieval starts with a user making types of arguments. i.e. “find pictures of Sachin Tendulkar” . User uses the possible range of color , texture, shape to express his or her query. image retrieval requires human feedback in order to identity higher-level concepts. SEMANTIC RETRIEVAL
- 10. 10 Feedback from users as “relevant” “not relevant” or “neutral” . Repeat the search with the new information. Using this, CBIR successfully understand the user intent. RELEVANCE FEEDBACK
- 11. 11 images with same color histograms Histogram determines the color quantization. The comparison between images is accomplished through the distance or similarity between the two histograms. Figure shows two similar images according to histograms. Feature Extraction Based On Color
- 12. Feature Extraction color layout 12 divide the whole image into a set of sub images. takes the color histogram of each sub image and global image. Using these histograms we can Mention the color auto, how the colors change with distance.
- 13. Feature Extraction Based On Shapes 13 Shapes Extraction includes edges, contours. Methods based on quantifying edge applies on query image. k-dimensional metric is used to match two images. Skeletal representations of object shape can also be used. For image matching suggested the use of directional histograms of the extracted edges.
- 14. Feature Extraction Based On Texture Texture refers to the visual patterns that have properties of homogeneity or arrangement. Calculate the relative brightness of selected pairs of pixels from each image(Query and stored images). The system then retrieves images with texture measures most similar in value to the query. 14
- 15. Image Indexing This is achieved by assigning descriptive metadata in the form of keywords. This metadata can be used for retrieval keys. Allow the user to add descriptors. 15
- 16. Matching Query to Image Feature Vector Representation: We can represent the features of an image by vector. Using the various formulas we can calculate the similarity between two images. Fuzzy reasoning can be implemented for matching the images. Query and User’s refinement.(region of the interest.) 16
- 17. Content Based Video Retrieval Video Parsing • Manipulation of whole video for breakdown into key frames. Video Indexing • Features of these frames are extracted and indexed based on color, shape, texture (as in image retrieval) Video Retrieval and browsing • Users access the database through queries or through interactions. 17
- 18. Applications Crime prevention The military Intellectual property Architectural and engineering design Fashion and interior design Web Searching Journalism and advertising Medical diagnosis Geographical information and remote sensing systems Cultural heritage Education and training 18
- 19. QUERIES?? 19
- 20. Thank You 20