Emotion Detection from Speech Using Machine Learning and Neural Networks
An algorithmic approach for detection and analysis of human (7 different) emotions with the help of speech processing, resulting in 80% efficiency with the use of Continuous Hidden Markov Model
View Project
Automatic Speech Recognition (ASR) Using Machine Learning
Detected speech from MFCC features Using Hidden Markov Model
View Project
Sudoku Solver Using Java and Android
Android Application which uses Backtracking Algorithm to Solve any 9*9 Sudoku Problem
View Project
Peer to Peer File Sharing Software Using Java
• Implemented P2P File Sharing Software, a simplified version of the Bit Torrent
• Main Features are Choking, Unchoking and Preferred Neighboring Selection Mechanism. Tested using a minimum of 6 peers and a large file of size greater than 1 Gb
Smart Stick for Visually Impaired Users Using IOT and Android
• Helped Blind Users to navigate to new locations outside their home by vibrating motors
• Implemented voice control module where user can input destination through voice, used Google Places API to autocomplete what location the user is looking for.
• User can also send their location for emergency alert to their family members
Encoding and Decoding Using HuffmanCoding in Java
Implemented Huffman Coding using 3 different data structures and compare the performance; Binary Heap, 4-way cache optimized heap and pairing heap.
View Project
Document classification Using Machine Learning
An algorithmic approach for Document classification model. Approached this Problem with Random Forest, SVM, Naive Bayes, LogisticRegression Classifiers and found out RFC to be the best Classifier. Then used Random Classification Model with Best Paramaters, resulting in 87% effeciency.
View Project
Alumni Network Using SQL, Python (Flask), Knockout JS and Android
Developed a social network platform of Alumni, Features included Viewing profile and Sending References.
A full-fledged Android Application was developed and connected to the backend
Craigslist Revamp Using React, HTML and CSS
• User Interface of “Craigslist” was made better by revamping the color scheme and the layout of website
• A more user-friendly search filter was made using sliders and drop-down menus
• Recommendation System, Wishlist and Social Authentication features were implemented