Georgia Tech's Computational Photography
Portfolio
Esaias [email protected]
https://cs6475.wordpress.com/fall-2016/
Assignment #1: A Photograph is a Photograph
“Spring in Michigan”Rochester Hills, MI, USAMay 6th, 2016 7:26PM
● Samsung-SM-G930A (Galaxy S7)● 1/530● f/1.7● ISO-50
Assignment #2: Image I/O [1/2]
Original image Grayscale Black and White
Assignment #2: Image I/O [2/2]
Input 1
Input 2
Overflow Output image
Correct Output imageInput 2 Grayscale
Input 1 Grayscale
Assignment #3: Epsilon Photography [1/4]
.
These photos represent my last name by using a small flashlight.
Final Artifact
Image 1 Image 2 Image 3 Image 4
Assignment #3: Epsilon Photography [2/4]● Each picture was taken with the same camera parameters and with a 30s delay.● ISO-100, f/25, 1/10 sec, 200mm
Assignment #3: Epsilon Photography [3/4]
● There are 2 final artifacts generated from the 8 pictures of the moon:○ GIF○ Blended image that shows the trajectory of the moon
over a 4min window (8 pictures taken 30s apart)
Assignment #3: Epsilon Photography [4/4]
Static Image: https://drive.google.com/open?id=0B9pB2VafJlaiYnYwajNyaTBROVU GIF: https://drive.google.com/file/d/0B9pB2VafJlaidzRNMUZkd2Y5dkU/view?usp=sharing
Assignment #4: Gradients and Edges [1/2]
For this assignment, I wanted to use a picture of a road so get a glimpse of how easy/hard it would be to build a “lane detection” algorithm
Original Own edge detection method Canny edge detection
Assignment #4: Gradients and Edges [2/2]
Original Own edge detection method Canny edge detection
Assignment #5: Camera Obscura [1/3]
The Scene The Image The Setup The focal plane
Assignment #5: Camera Obscura [2/3]
These are 3 pictures covering the entire Field of View of the camera obscura
Assignment #5: Camera Obscura [3/3]Using the pictures from the last slide, created a panorama with Hugin and exported it to TIFF then using GIMP I flipped vertically to obtain the following artifact (in JPG)
Assignment #6: Blending [1/2]
Black WhiteMask
● What is the subject of your images? Are the images your own? (If not, cite your source.)○ Both of them are images I captured using my smartphone
● How did you create your mask? (Remember to mention any tools you use.)○ I opened the “white” image in MS Paint and drew a white rectangle where the SD is and
Filled with black around it.
Assignment #6: Blending [2/2]
Final Blended Image
● Tried different sets to show the behavior of the ORB algorithm using:○ Objects that cause specular reflection○ Textured objects○ Objects that cause diffuse reflection
● Had to make changes to the provided code to get better results
Assignment #7: Feature Detection [1/4]
Assignment #7: Feature Detection [2/4]
Sample Lighting
Rotation Scale
Match Set Correct Matches
Notes
Sample 10
Lighting 8
Rotation 5
Scale 0
Set using object with specular reflection
Assignment #7: Feature Detection [3/4]
Sample Lighting
Rotation Scale
Match Set Correct Matches
Notes
Sample 1
Lighting 5
Rotation 1
Scale 3
Set using textured object
Assignment #7: Feature Detection [4/4]
Sample Lighting
Rotation Scale
Match Set Correct Matches
Notes
Sample 9
Lighting 9
Rotation 10
Scale 9
Best performing set, using object with diffusive reflection
Assignment #8: Panoramas [1/2]
Sample Input
Images
Output
Assignment #8: Panoramas [2/2]
Own Input
Images
Output
Assignment #9: HDR [1/4]Sample Input images
Assignment #9: HDR [2/4]
Output
Assignment #9: HDR [3/4]
1/5 1/13 1/20
1/40 1/100 1/160
Assignment #9: HDR [4/4]
Output
Assignment #10: Pictures of Space [1/4]
Link to original pictures: https://drive.google.com/open?id=0B9pB2VafJlaiSlVXdHF2S0lpaDg
Assignment #10: Pictures of Space [2/4]
https://photosynth.net/view.aspx?cid=5731f2b4-8bc7-4d47-ad1f-c8fe396593df&m=false&i=0:0:4&c=-6.22913:1.1478:-0.108385&z=652.28835002729&d=2.56415218185859:-2.48889958600817:-1.86771307948356&p=0:0&t=False
Assignment #10: Pictures of Space [3/4]
Selected images of the panorama
Link to original pictures: https://drive.google.com/open?id=0B9pB2VafJlaicXdpdnkyVnBsc0E
Assignment #10: Pictures of Space [4/4]
https://photosynth.net/view.aspx?cid=fd1e621c-6241-4a4c-8908-6e4b673e6e7f
Assignment #11: Video Textures [1/2]
● Candle video texture gif: ○ https://drive.google.com/open?id=0B9pB2VafJlaiSzhqZC1FNGJiLUU
Start Frame End Frame
Start frame: 40File name: candle 040.pngCode index: 0
End frame: 92File name: candle 092.pngCode index: 52
Assignment #11: Video Textures [2/2]
● This my finger typing both the ‘s’ and the ‘d’ keys on my laptop’s keyboard.● Link to your video texture gif:
○ https://drive.google.com/open?id=0B9pB2VafJlaiU3NDYmxvSm1mMmM● Link to either the frames (folder) or the video clip you started with:
○ https://drive.google.com/open?id=0B9pB2VafJlaiR0paVDRwZlozV3M
Start Frame
Frame: 43Name: frame0043.pngIndex: 40
End Frame
Frame: 169Name: frame0169.pngIndex: 166
Final ProjectSummaryA Surround View System allows a driver to have a “bird’s eye” view of the surrounding environment around a vehicle. This allows for increased safety while operating it - e.g. parking a car and/or maneuvering in tight spaces.
GoalsUse the video input from 4 fisheye lens cameras with 180 degree FOV, one on each side of a vehicle, and create artificial bird's eye view of the vehicle surroundings. This is a parking assist feature available in some high-end cars nowadays.The intention is to be able to work with both static images and live video, captured from cameras in real-time.TeamKaloyan Tenchov <[email protected]> & Esaias Pech <[email protected]>Example image setshttps://drive.google.com/drive/folders/0B6E0LnuGgjxKWTAzbGtublhNSXM
Final Project - Examples 1/3Input Output
Final Project - Examples 2/3Input Output
Final Project - Examples 3/3Input Output
Final Project - Pipeline 1/3Fisheye
Equidistant Projection
Fisheye Equidistant Projection
Fisheye Equidistant Projection
Fisheye Equidistant Projection
Front Camera Input
Right Camera Input
Back Camera Input
Left Camera Input
Bird-eyePerspective Projection
Bird-eyePerspective Projection
Bird-eye viewPerspectiveProjection
Bird-eye viewPerspectiveProjection
Rotate 90°
Rotate 180°
Rotate 270°
CombinedImages
CroppedSurround
ViewImage
Final Project - Pipeline 2/3Front
Right
Back
Left
Final Project - Pipeline 3/3
Surround View
Final Project - Custom Camera Rig
Custom fisheye camera rig, constructed out of plastic picture frames, and binder clips, specifically for the project. The cameras used are ELP-USBFHD01M-L180 from Amazon.