Pranav Srivastava

Hello World

• Homepage
• Medium
• Github

• Etiam posuere

• 01

  identifying feature points and image matching

  Uses concepts of Machine Vision. Has 2 parts: (A) identifying interest points in an image together with their scales and rotations (method inspired by SIFT algorithm) (B) find a patch extracted from one image in a second image using correlation-based area matching algorithm

  explanation

• 02

  Scene reconstruction

  Checkerboards are used for calculating calibration matrix. Further, feature points in the frames are refined to sub-pixel accuracy using the calibration matrix. These feature points are tracked across the video frames using KLT algorithm. Feature points common in first and last frame of the video are projected, later the errors are also projected.

  demo

• 03

  Chatbot with a personality and short term memory

  This chabot is built on the top of a baseline chatbot model from Stanford. Have used the core concepts from NLP. The additions made are: (A) trained on multiple datasets (B) it has three personalities (C) it can remember the information from the current conversation (D) it can be trained by the user, on the fly

  demo

• 04

  Get context from the conversation

  This project is to demonstrate how to use speech2text API from Contexta, in your app, for getting the context.

  code

• 05

  Chatbot using SMS API

  This project is to demonstrate the concept of messaging. Further, this project aims at building a chatbot using SMS API and intent classification.

  code

  web version

• 06

  Genetic Algorithm for TSP

  This project is the implementation of search heuristic for TSP (Traveling Salesman Problem) using Genetic Algorithm. The end goal is to find the least costly route for the salesman. TSP is representative of combinatorial optimization problems known as NP-hard.

  code

• 07

  Build a range of deep learning models using convolutional neural networks

  This is a multi-class classification problem with 17 possible classes (17 different classes of flowers). It is focused on data augmentation, ensemble techniques and transfer learning.

  code

• 08

  Headpose and eyegaze tracking using OpenFace 2.0

  This project focuses on the setup on MacOs and tracking headpose and eye gaze.

  SetupTutorial

  demo