From 9a1afc10d42d24147398cd759de88cb4797223fd Mon Sep 17 00:00:00 2001 From: Navreet Kaur Date: Thu, 14 Nov 2019 23:18:10 +0530 Subject: [PATCH] Update README.md --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index d9b714d..df83056 100644 --- a/README.md +++ b/README.md @@ -4,6 +4,7 @@ Assignment 2, Deep Learning (ELL888) @ IIT Delhi ## Problem Statement: Lots of good quality educational videos are available on platforms like YouTube and NPTEL but students are not able to use them optimally since videos are not indexed properly. This assignment tries to deal with this problem. Have a look at this video: https://goo.gl/VTaz7h + You will notice that, in general, the professor writes on papers, removes that paper that is completely filled and then starts writing again on a new sheet. The aim of this assignment is to extract out those frame of a video which captures the full written content on a sheet of paper. Suppose the professor writes 7 sheets in one lecture, if we could get the frames capturing those 7 filled sheets, we will have almost all the information of what professor taught in that lecture in just 7 images. One hour video can be summarised using 7 images (Call these 7 frames as ‘key-frames’). Sometimes, instead of handwritten sheets, other mediums are used to teach in lectures (ppts, notepad etc), but the core problem is still the same,