Rudrax Dave's repositories
LostNFound
Lost-and-Found Application for University Operations -Solved problem of Lost or Recovered items with features such as image uploads, image recognition, and smart matching with technologies like Android Studio, Google Firebase- Storage, Authentication, Realtime Database, and Google Analytics • Built an interface for mobile application will let people report when found or lost something at a specific concerned location allowing users to capture and upload files into the system cloud and save them in the database, can be automatically matched to images in the database- and shown to users in the recovery section of the app.
UrbanizationDetection_RoadnBuilding
Building and Road Extraction for Urban and Rural Development and Annotations of Imagery -Jan 2021 - Jun 2021 Associated with Bhaskaracharya Institute For Space Applications and Geo-Informatics -Utilizing Open-Source Datasets from Google Earth Engine & NASA USGS (Sentinel, Landsat-8) of 2 certain timestamps, equalizing tiff files by Histogram Eq. Method, Clustering data by PCA + K-means Methodology, trained and segmented Data by Deep Learning Algorithms with U-NET Architecture, computed results by confusion matrix and attaining accuracy 89 percentage. • For mapping from high resolution imagery or GIS database construction and its update, automatic object-based image analysis, also animated change- after Change Detection Model so users come to know how urbanization occurs or growth happens over a decade.
Gene_Disease_Association_Prediction_with_GAT
This project focuses on analyzing and predicting Gene-Disease Associations (GDA) using graph-based machine learning techniques. It leverages curated datasets, protein-protein interaction (PPI) data, and various node features to construct graph representations of gene-disease associations. Two graph neural network architectures, GraphSAGE and GAT.
AmericanSignLanguage_Reader
Project Data: The Dataset is 1.11 GB in Size. The images in the dataset are manually captured and not computer-generated. The dataset linked above contains images from 29 classes (26 alphabets, SPACE, DELETE, and NOTHING). Each class contains 3000 images in the training set and each image is a 200 x 200 RGB image. The training data set contains 87,000 images, of which 26 are for the letters A-Z and 3 classes for SPACE, DELETE, and NOTHING. These 3 classes are very helpful in real-time applications and classification. The test data set contains a mere 29 images, to encourage the use of real-world test images.>the test set is very small. We plan to Experiment with different architectures and hyperparameters. First, implement the CNN architecture described in the Models of [8] and apply our training data to generate the classification of 26 Letters. Our goal is to map images from a particular domain to a Letter that it means. We will be finetuning (transfer learning) existing models that perform classification on RGB images. We will Look at the levels of RGB images in Particular Centroidal Pixels and We can Classify from the Centre Which Shape is Defined. We are Planing to Use RESNET - Residual Neural Networks or Similar to be able to argue and analyze the better Accuracy between all of the Networks.
BeaconNotify
This embedded app, allows user to get notifications for a particular designed content when they are in range of a beacon! As Nearby notification and Physical web have been terminated by Google, we needed an external app to get notification from beacons. I've placed the screenshot of the notification when you are in range, out of the range and th…
CampusNavigation
final-project-RudraxDave created by GitHub Classroom - This project focuses on using data structures in C++ and implementing various graph algorithms to build a map application.
EE538
Repo for EE538
ForestFires_Prediction
This dataset contains weather data from 2 regions in Algeria over the period of 3 months and the goal is to predict if a fire occurred at any day within that period. To create a real-world scenario, we want to predict if there will be a fire in a future date as provided by the dataset. The fire prediction is based on weather data collected from the regions. Problem type: Classification.
LA_Foreclosure_Rates_Analysis
Incremental Analysis on Foreclosure Rates for City of Los Angeles