Skip to content

A web app that uses a relatively small fully convolutional neural network to brighten low light images.

Notifications You must be signed in to change notification settings

range123/SeeInDark

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

See In The Dark

This project is used to brighten dark (low light) images and allow us to observe the objects and details in the low light images. Uses a fully convolutional neural network (contains convolution and transposed convolution layers) for brightening the images. This repo contains a flask app where the model is deployed. The model is trained on the "See In The dark" sony dataset and based on Learning-to-See-in-the-Dark.

Key Features

  • The network is relatively small and runs interactively on edge devices.
  • Works well on even almost completely dark images.

Prerequisites

  • Python3.6+
  • pip3/conda

Installation

  • Clone this repository.
  • (optional) Create a virtual environment.
  • Install the requirements.
pip install -r requirements.txt
  • Run the flask app
export FLASK_APP=app.py
flask run

Screenshots

MainPage Results

About

A web app that uses a relatively small fully convolutional neural network to brighten low light images.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 51.4%
  • HTML 47.7%
  • CSS 0.9%