Video-Super-Resolution

Om Godage

---

Things that I've learnt during SoC:

1. Some basics of Machine Learning
   1. Principal Component Analysis
   2. Supervised and Unsupervised Learning
   3. Linear regression
   4. Logisitic regression
   5. Classification (Multiclass)
   6. Neural Networks
   7. Convolutional Neural networks
2. Some basics of Image processing
   1. Blur kernels
   2. Convolution
   3. PSNR
   4. Conversion of RGB images to YCbCr, Y and others
3. Miscellaneous
   1. Markdown
   2. Using OpenCV to extract frames from a video
   3. Git and version control
   4. Some basics of vim

---

Week 1: In the first week of the SoC, we learnt about principal component analysis (PCA), brushed upon python and terminal commands. We applied PCA on the MNIST dataset and compressed the dataset to 10% of it's original size, without losing much data
Week 2: We learnt about linear and logistic regression, learnt aobut supervised and unsupervised learning, finished the first 4 weeks of the Andrew Ng course on Machine Learning, wrote the very first machine learning code to identify handwritten digits from the MNIST dataset.
Week 3: We learnt about Neural Networks, and wrote a program to identify Cats and Dogs from their images.
Week 4+5: We learnt about convolutional Neural Networks, And some basic image processing. We were introduced to the actual problem statement here.
Week 6+7+8: We wrote the code for the actual problem statement, first started with Image Super Resolution, then extending it to Frame-wise video Super Resolution. We tested it on multiple videos and documented our results, finishing off with some final touches