Manual annotation of Aerial Images is extremely time consuming.
In this work, we follow the idea of Achilles is to produce polygonal annotations of objects interactively using humans-in-the-loop.
Development Tracker:- This module performs a time series analysis on a set of segmented images. For eg: we will be providing two images one which is the present image of a particular area and another will be the image of the same area but 10 years back, by doing so we will easily track down how much growth has been done in that particular area in terms of infrastructural development, deforestation, etc.
QGIS Support:- This module takes a segmented image as an input and returns various files through which users can add vector layers of different segmented classes in various GIS softwares. Mainly for QGIS as it is the tool used by the institute but we have also verified it’s functionalities for ArcGIS.
LabelMe Support:- LabelMe is also an open-source available tool which helps people annotate data manually. This support helps the user in correcting some errors that might occasionally occur results generated by our deep learning techniques. It saves a lot of time as the user doesn't have to label the entire image instead he/she just have to make amendments.