Computaional Photography portfolio

  • View
    243

  • Download
    0

  • Category

    Design

Preview:

Citation preview

Georgia Tech'sComputational Photography

Portfolio

Apurva Guptaapurva.gupta@gatech.edu

PHOTO-MOSAICnot ‘Pixel Perfect’, but ‘Picture Perfect’

Final Project

Slice Source image into tiles

Resize images from the image corpus keeping their aspect ratio

Find average brightness of each color channel of the corpus images

Compare the distance of the average brightness of each image with the average brightness of source image.

Replace each pixel in the source image with the image with the least distance

PHOTO-MOSAICnot ‘Pixel Perfect’, but ‘Picture Perfect’

Average Brightness

Median Brightness

Assignment #1A Photograph is a Photograph

Camera: NIKON COOLPIX S3100F-Stop: f/5.5

Exposure time: 1/1000 secISO speed: ISO-80

Focal length: 18mmExposure bias: 0 step

Max aperture: 3.4Metering Mode: Pattern

Flash Mode: No FlashDimension: 3240 X 4320

The Capitol

Assignment #2

Black and White

Used nested loop to traverse each pixel and checked the threshold value of 128. Value greater that 128 corresponded to 255 while the smaller one was changed to 0.

Horizontal Flip

Used nested loop to traverse each column in the image matrix. The (x,y) pixel values in each column was replaced with the value of (x, height-y) pixel value and stored in another 2D array.

Average of two images

Used nested loop to traverse the images and added the (x,y)th pixels of both the images and divided the values by 2. This was stored in a 2D array.

Assignment #3Epsilon Photography

Epsilon Photography

Ghosts on Road

Assignment #4Camera Obscura

The Setup

Pin Hole was created by covering the window.Pinhole

The Image

This wasn’t visible through eye but came up well on camera. The edges are not sharp since it was a big pin hole. The building specifics were not visible at all but the sky came out really good.

The Image

Assignment #5Gradients and Edges

X Gradient

For the X gradient, I subtracted the ith pixel from i+1th pixel looping through the columns.

For the Y gradient, I subtracted the ith pixel from i+1th pixel looping through the rows.

Y Gradient

Original image

Kernel image

Black and white imageThreshold - 100

Original image

Kernel image

Black and white imageThreshold - 150

Edge Detection

Assignment #6Blending

Output

Black

White

Mask

BlackWhite

Mask

Output Output

When converting the output to grayscale, it looks as It there is shadow of tree on the surface .

To create this mask, I took the Black image and using the threshold of 128 and changed the pixel value to 0 or 255

Black White Mask Output

Similarly I created masks for black image and tried to blend it with the white image.

I tried these with various threshold values to create masks and check the resultant blending.

Mask OutputMask Output

Assignment #7Feature Detection and Matching

Score: 3/10

Lighting

Score: 2/10

Rotation

Score: 4/10

Sample

Score: 10/10

Scale

Assignment #8Panorama

Sample Panorama created using inbuilt Phone feature

Assignment #9Photos of Space

PhotosynthFord Museum + Campionile

PanoramaCampionile

PhotosphereCampionile + Labspace

Assignment #10HDR

Output with the given set of images

Output with the given set of images on the right.

Output with the given set of images on the right.

Assignment #11Video Textures

Output Link:https://drive.google.com/file/d/0B5Ncl02d4dOeS2JxeVIycEVMdmc/view?usp=sharing

https://drive.google.com/file/d/0B5Ncl02d4dOeNjVXS3BEa0pLeFU/view?usp=sharing