Kauê de Moraes Vestena's repositories
opensidewalkmap_beta
OSWM is intended to be a full-fledged inventory project for pedestrian networks management using OSM data and fully hosted and updated using github
Curitiba_OSM_Importing_Codebase
This repository holds the code for the importation of the data from the "Geocuritiba" Portal, available at
deep_pavements_runner
Main Functional Repository for the Deep Pavements Framework Project. It shall be the one used to process and acquire data by the end user.
oswm_codebase
This repository holds the code that is used to create the static data for each node of project OpenSidewalkMap, that is a decentered and modular project, leveraging OpenStreetMap data for sidewalk data management.
my_mappilary_api
my personal one, less dependencies
deep_pavements_dataset
This repository holds the dataset to be used to improve semantic models on the classification of pavements. All images came from Wikimedia Commons, Mapillary or Flickr, licensed under CC Compatible licenses.
deep_pavements_sample_picker
Repository to leverage the capacities of recent semantic segmentation models with language models, and train a more "conventional" model to identify roads & sidewalks alongside pavement type
opensidewalkmap-1
Easily visualize sidewalks
htmpa
H.T.M.P.A = Hot Tasking Manager Project Aggregator. The idea is to be a template repository to keep a list of tasking manager projects related to a theme to simplify project finding. It's gonna be updated automatically daily using GH actions, maybe even a telegram bot for updates...
osm_sidewalkreator
Plugin designated to create the Geometries of Sidewalks (separated from streets) based on OpenStreetMap Streets, given a bounding polygon, outputting to JOSM format. It is mostly intended for acessibility Mapping.
deep_pavements_project
This repository is meant to be the main one to organize the project Deep Pavements, which is meant to leverage AI tools to do pavement surface segmentation in CC-Compatible street-level imagery, with categories types complying with OpenStreetMap Tags.
3d_patches_sandbox
This repository is a draft to start a new project based on combining Semantic Segmentation and Monodepth to extract 3D features using Mapillary Imagery
deep_pavements_trainer
Repository to hold the scripts to train models for the Deep Pavements project
lang-segment-anything
SAM with text prompt
kauevestena.github.io
My personal webpage
open_llama
OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset
deep-image-matching
Multiview matching with deep-learning and hand-crafted local features for COLMAP. Supports high-resolution formats and images with rotations. Both CLI and GUI are supported.
editor-layer-index
A unified layer index for OSM editors.
deep_pavements_labeler
The part of project Deep Pavements were the samples collected using the Sample Picker are prepared to become a training dataset.
CLIP
CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image
test_tiles
A test repository for testing options for generating webmaps with vector tiles from data obtained from OpenStreetMap
opensidewalkmap
Repo to host OSWM index map
monodepth_tests
a repository to test different monodepth algorithms and approaches, aiming to find a working and functional one among the ones with compatible licensing
sidewalk_analysis
A repo to host some scripts for some analysis, maybe it can become a tool... Webmap of the analyzed data on the paper in the link:
mds_leonardo_moreira
parte da colaboração no trabalho de doutorado do Leonardo Moreira