I have face images which will be copied into a folder dynamically. I need to remove duplicate faces and show only distinct faces.
approach 1) Initially i tried using LBP approach in which i constructed histogram for every image and copied the imagepath/histogram as key/value in hashtable. started comparing with other image histograms which is cpu intensive operation i.e if i have 1000 histograms i need to compare histogram 1 to all remaining histograms. Finally this approach was not solving my problem. I was stuck here
approach 2) If i can group all similar face images into specific cluster and later remove the duplicates iteratively. Please let me know using python how can i use kmeans for unlabelled data.
Thanks cd
