ELEC474 Computer Vision Project
Advisor: Michael Greenspan
(Notion Link)
Python with OpenCv, just check the code in the ./lab1
Prelab2 ---- Flood Fill and Homogeneity criterion correct
Original
Tensity= 52
Tensity= 103
Lab2 ---- K-Means and Color Quantization
Original
K value = 2
K value = 6
Prelab3 ----- General and Canny Edge Detection
Original
general t = 51, Canny t = [51, 102]
general t = 21, Canny t = [30, 7]
Implement different method for sorting and searching with different data structure. The best solution is ./lab3/Main_MikeFeng_20119641_Gresult_F.ipynb
Original
With in 60s for 100000
Prelab4 -- BFMatch with Lowe Ratio
------
Original imgs
------
descriptors
Matches
General and low ratio
Lab4 ---- Transformation Estimation
descriptors
Affine Transform apply
Perspective Transform apply
Prelab5 ---- Epipolar Geometry
Epipolar Calculation
Point/Conjugate Epipolar Line Selection
Lab5 ---- Stereo Rectification and Block Matching Disparity Map Calculation
Stereo Rectification original(l)
Stereo Rectification
Block matching original (l)
Block matching
The project has using the imgs in ./Project/offices2
Detail can check the report as .docx file or ./Project/Deliverable/D2/Project_Report.pdf
Stich version 1
Stich version 2 (best)
Stich version 2
Stich version 3