Submitted Final Presentation

ShrayanRoy · May 22, 2024 · be6e11e · be6e11e
1 parent 5fdeb0e
commit be6e11e
Show file tree

Hide file tree

Showing 10 changed files with 1,359 additions and 10 deletions.
diff --git a/presentation/Endterm_presentation.Rmd b/presentation/Endterm_presentation.Rmd
diff --git a/presentation/Submitted-Endsem presentation.html b/presentation/Submitted-Endsem presentation.html
diff --git a/presentation/finalpresentation.Rmd b/presentation/finalpresentation.Rmd
@@ -42,9 +42,9 @@ xaringanExtra::use_panelset()
 
 # Depth: the third dimension
 
-* Traditional photographs are two dimensional projections of a three dimensional scene.
+* We want to estimate **depth**, which represents the distance between camera and objects in the image.
 
-* The third dimension is **depth**, which represents the distance between camera and objects in the image.
+* Depth estimation from single image is more challenging, as we have only one observation per pixel. 
 
 ```{r ,warning=FALSE,echo=FALSE,out.width='70%',fig.align='center',echo=FALSE,fig.cap= "Figure: 3D perspective from 2D image"}
 
@@ -188,11 +188,11 @@ knitr::include_graphics("pimg/7.1.png")
 
 * The entire procedure is time-consuming, even the initial blur map estimation also.
 
-* Boundary between different objects in image is not clear.
+* Boundary between different objects in image may not be clear.
 
 --
 
-* Instead of estimating blur for all pixels separately, consider the individual objects.
+* Instead of estimating blur for each pixels separately, consider only the individual objects.
 
 * It is the objects themselves that vary in distance from camera, not individual pixels.
 
@@ -539,7 +539,7 @@ knitr::include_graphics("pimg/refocus1.png")
 
 * Our R implementation takes around 7 minutes time for estimation step for $1000\times1000$ image.
 
-* SAM takes around 3 minutes for segmentation on GPU run time.
+* SAM takes around 3-5 minutes for segmentation on GPU run time.
 
 * Pixel by pixel blur estimation takes around 1 hr 30 minutes only for initial step.
 

diff --git a/presentation/finalpresentation.html b/presentation/finalpresentation.html
@@ -52,9 +52,9 @@
 
 # Depth: the third dimension
 
-* Traditional photographs are two dimensional projections of a three dimensional scene.
+* We want to estimate **depth**, which represents the distance between camera and objects in the image.
 
-* The third dimension is **depth**, which represents the distance between camera and objects in the image.
+* Depth estimation from single image is more challenging, as we have only one observation per pixel. 
 
 &lt;div class="figure" style="text-align: center"&gt;
 &lt;img src="pimg/dof.jpg" alt="Figure: 3D perspective from 2D image" width="70%" /&gt;
@@ -186,11 +186,11 @@
 
 * The entire procedure is time-consuming, even the initial blur map estimation also.
 
-* Boundary between different objects in image is not clear.
+* Boundary between different objects in image may not be clear.
 
 --
 
-* Instead of estimating blur for all pixels separately, consider the individual objects.
+* Instead of estimating blur for each pixels separately, consider only the individual objects.
 
 * It is the objects themselves that vary in distance from camera, not individual pixels.
 
@@ -513,7 +513,7 @@
 
 * Our R implementation takes around 7 minutes time for estimation step for `\(1000\times1000\)` image.
 
-* SAM takes around 3 minutes for segmentation on GPU run time.
+* SAM takes around 3-5 minutes for segmentation on GPU run time.
 
 * Pixel by pixel blur estimation takes around 1 hr 30 minutes only for initial step.
 

diff --git a/presentation/pimg/dest.png b/presentation/pimg/dest.png
diff --git a/presentation/pimg/reconstr1.png b/presentation/pimg/reconstr1.png
diff --git a/presentation/pimg/reconstr2.png b/presentation/pimg/reconstr2.png
diff --git a/presentation/pimg/reconstr3.png b/presentation/pimg/reconstr3.png
diff --git a/presentation/pimg/reconstr4.png b/presentation/pimg/reconstr4.png
diff --git a/presentation/pimg/rest.png b/presentation/pimg/rest.png