Conv Nets IV

In this project, I extend my CNN Advanced Architecture project to 3D images. The architecture is advanced in the sense that it is flexible and can read & build its architecture from JSON files. The 3D extension allows CNN3D classifier to tackle 3D tensors used heavily in scientific/medical imaging such as CT Scan images, which we will we working with in this repo. I have trained the LUNA16 dataset with the usual my CNN and then with CNN3D classes and noted that CNN3D outperfromed the predictions for...

Conv Nets II

In this project we automate learning CNN hyper-params. We test and improve my previous tensorflow implementation of CNN in here on Kaggle's facial expression dataset. Here I only deal with two label classes: Happy and Angry. Randomized Search for learning hyper_paramters such as learning_rate, momentum coefficient, etc and drop-out regularizer improved the default test prediction accuracy from 36% to 64%. This is a bit lower than 72% that we got from our SVM classifier. However, we will continue in following projects to see if we can beat...

Conv Nets I

This project implements CNN class using the tensorflow library. I compare the accuracies gained by the ConvPool layers to the ANN architectures of the ANN-tensorflow repo. We find that adding only 2 ConvPool layers before the ANN architecture does automated feature engineering and significantly increases the classification accuracies as compared to the ANN only architectures....

Image Restoration

This project tackles image restoration and coloring for old damaged black and white images. The restoration problem tackels automatic damaged region detection which turns out to be a hard problem to automate. The successful resotrations still require some form of tweaking the Kernel parameters (dilate/erode/color threshold, etc.) in the code. The coloring portion Opencv DNN based Caffe framework to colorize black and white images. One can find the Caffe models here...

Camera Calibration

In this project, I use simple manual camera calibation method. The setup is as follows: Assuming pin hole camera, we can utilize similar triangle methods(see below), and can come up with focal lengths equation: # fx = (dx/dX) * dZ , fy = (dy/dY) * dZ where dX, dY are the physical length & width of the object in view and dZ is the distance from object to camera. dx and dy are the corresponding pixel width & height of the object...

Features Matching II & Homographies

This project implements feature generation and matching using OpenCV library. We observe two classes of object types: The first is solid cube shaped object such as a cereal box that is ideal for homography transforms and post processes. The second is deformable object such as a bag of chips that can be physically distored between two images and homographies and post processing may fail on them. In this project we observe the usage of optical flow on the 2nd category to check viability of estimating such distortions....