https://github.com/narendrakumar92/Video_Feature_Analysis

Detects and extracts the features such as color, key points, and motion vectors. Helps in finding Video similarity, Video subsequence search, Dimensionality reduction, Querying, Indexing (PageRank and ASCOS++), Multidimensional Index Structure and KNN (Locality Sensitive Hashing tool), Visualizing the query result as videos
https://github.com/narendrakumar92/Video_Feature_Analysis

Last synced: about 2 months ago
JSON representation

Detects and extracts the features such as color, key points, and motion vectors. Helps in finding Video similarity, Video subsequence search, Dimensionality reduction, Querying, Indexing (PageRank and ASCOS++), Multidimensional Index Structure and KNN (Locality Sensitive Hashing tool), Visualizing the query result as videos

• Host: GitHub
• URL: https://github.com/narendrakumar92/Video_Feature_Analysis
• Owner: narendrakumar92
• Created: 2016-12-15T01:17:09.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2017-03-23T02:56:51.000Z (over 7 years ago)
• Last Synced: 2024-06-09T20:40:36.056Z (4 months ago)
• Language: Matlab
• Size: 2.92 MB
• Stars: 4
• Watchers: 1
• Forks: 5
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Video_Feature_Analysis

• We detect and extract the features such as color, SIFT key points and motion vectors in every frame of the video set. To do so, we use the libraries such as SIFT and FFmpeg to obtain the dataset. 

• We identify the color similarity, SIFT similarity and motion similarity between two video files using several distance similarity measures. Also, finding the overall similarity between the videos.

• We do Video Subsequence search, returns the 'k' most similar frame sequences, and visualizes the query and result frames as videos. We do this in the normal space as well as reduced space using PCA and K-means.

• We Find the most significant frames and content using PageRank and ASCOS algorithm to identify what the given set of videos majorly contains. 

• Object querying using Approximate algorithm, Locality Sensitive Hashing using random Hyperplanes to find the similar objects in the database and visualizing the query result as videos.  

Please read Phase1_Taskdetails.pdf , Phase2_Taskdetails.pdf and Phase3_Taskdetails.pdf for the complete details on each of the project phases.