Discover @Tejas1415 Open Source projects

Tejas Krishna Reddy (@Tejas1415)

Tejas1415

Stars
58
Global Rank 307,272 (Top 11 %)
Followers 33
Following 3
Registered over 6 years ago
Most used languages

MATLAB
40.9 %

Python
31.8 %

Jupyter Notebook
18.2 %

C
4.5 %

C++
4.5 %
Location 🇺🇸 United States
Country Total Rank 61,933
Country Ranking

MATLAB
695

Jupyter Notebook
8,639

Lidar-Based-SLAM

Robotics Sensing and Navigation, Localization and Mapping in Python using Open3D library.

Multiple-Moving-Object-Detecting-and-Tracking-Using-ML

Multiple Moving objects in a surveillance video were detected and tracked using ML models such as AdaBoosting. The obtained results were compared with the results from Kalman Filter.

Neural-Machine-Translation-NMT

Using Google Colab, we develop a NMT, language translator. Here, we do NMT to translate from English to Vietnamese.

Jupyter Notebook

Discreet-Cosine-Transform-DCT-Singular-Value-Decomposition-SVD-based-Copy-Move-Forgery-Detection

Implementing Copy Move Forgery Detection using DCT or SVD transformations. Also, a clear explanation is provided in the READme section

Copy-Move-Forgery-Detection-Using-Hu-s-Invariant-Moments

Through the proposed algorithm, the features extracted from the image can be reduced to as less as 1 feature per block. Paper Published in 3rd IEEE conference(RTEICT) 2018, SVCE college, Bangalore.

Automated-region-masking-of-latent-overlapped-fingerprints

The tedious manual process involved in the first step of overlapped fingerprint separation was completely automated using this algorithm. Published in IPACT, IEEE conference

Hackathon__Beer-Score-Prediction__Disparate-programs-using-Neural-Net__Convolutional-Neural-Net.

Regression Using NeuralNet and CNN and a detailed analysis over the results obtained from regression techniques in ML and DL.

Intensity-Brightness-Invariant-CMFD

An algorithm that is completely robust to Intensity/ Brightness varied copy move forgery is proposed in this algorithm.

Hu-s-Invariant-Moments-in-MATLAB

An improved and tested code to produce Hu's Invariant moments for any Image/ Audio signals. Hu's Invariant Moments are One of the Best Feature Extraction Techniques for Further Analysis.

Cifar10-image-classification-using-MLP-in-TF-backend.

Using TensorFlow backend, multiple methods and their results to achieve best classification for CIFAR10 image dataset. Edit: I have also included a complete keras guide (Colab Notebook) to build CNN-single Layer, CNN-Multi Layer and Transfer learning based CIFAR10 classification.

Jupyter Notebook

SIFT-Technique

Scale Invariant Feature Transform MATLAB code.

Image-Stitching-Panorama

From a set of images with overlapping content, we use Harris features to attach them all together into a single Panoramic Image

Machine-Learning-in-Cyber-Security

Course Projects at Northeastern University. Some really good projects are here with datasets.

Jupyter Notebook

Automatic-NPR-using-unique-Edge-Detection-Algorithm-and-Smarter-Analysis-Through-IoT

An improvised version of ANPR system has been implemented and a unique prototype was developed to avail maximum benefits to the society using ANPR systems using IoT protocols.

Rescale-Invariant-CMFD

For the first time in the research history of copy move forgery detection(CMFD), we propose an algorithm that can successfully detect copy-rescale-move forgery.

Web-Scrape-Stock-Ticker-and-Company-Name-Datasets

Different Stock Tickers and their corresponsidng company names are hard to find. So, I have written certain scripts which can be used to download the datasets

Genetic-Algorithm-PasswordHack

With a known answer the algorithm finds a way to reach the answer by itself. The code here is very basic and can be improved in a million ways but gives a clear essence of GA effect.

Task-Scheduling-in-Mobile-Environment

How different tasks are scheduled at different cores or cloud environment in a mobile environment based on the load of tasks and balancing a trade off between the fastest execution Vs Total Energy consumption to get an optimal result.

Power-of-Parallel-Processing

Kaggle competition to predict customer transactions was performed using Linear Regression, Neural Networks and Random Forests. Separate programs were written for each of the above models with and without parallelization and the time improvisation was noted.

Siamese-Networks-to-Classify-Real-Vs-Forged-Signatures

Using Siamese Neural networks to implement few shot learning process to achieve an amazing 97% accuracy in classifying real vs forged signatures.

Jupyter Notebook

Character-Recognition-Through-Deep-Learning

Basic Deep Learning character recognition model which could be used in various OCR required situations.

House_Price_Prediction_Hackathon

"3rd PRIZE WINNER" Hackathon: Predicted price of the houses located in Bangalore, India - over a real time large scale data provided by the hackathon team.