Code Monkey home page Code Monkey logo

mask_rcnn's Introduction

Mask R-CNN for Object Detection and Instance Segmentation

This is an implementation of Mask R-CNN on Python 3, Keras, and TensorFlow. The model generates bounding boxes and segmentation masks for each instance of an object in the image. It's based on Feature Pyramid Network (FPN) and a ResNet101 backbone.

This repository adds functionalities to the original repo to train on rgbd data for indoor scenes. In this case the SUNRGBD dataset is used to train the model and perform instance segmentation on the following classes:

  • desk
  • chair
  • sofa
  • bed
  • bookcase

Qualitative results

input data image 1 image 2
RGBD
RGB

Quantitative results

Functionality to calculate those metrics out of the box are added to the evaluation command. To execute the per class evaluation, the weights must be put into the logs/reduced_classes folder and the evaluation flag must be provided when starting the sun.py script. The weighted mAP among classes is 4.6 % higher when using rgb data what can be seen int the tables below.

Augm. number Augm. strength mAP (RGB) mAP (RGBD)
0 0 0.252 0.264
2 3 0.237 0.296
3 7 0.226 0.288

The mIoU and mAP for the best rgb and the best rgbd models for each class can be seen in the table below.

class mAP best RGB mAP best RGBD mIoU best RGB mIoU best RGBD
bed 0.331 0.285 0.351 0.295
chair 0.367 0.40 0.453 0.475
table 0.258 0.335 0.316 0.377
sofa 0.132 0.147 0.183 0.2
bookcase 0.100 0.143 0.141 0.189

mask_rcnn's People

Contributors

waleedka avatar yannick947 avatar keineahnung2345 avatar philferriere avatar borda avatar coreyhu avatar cpruce avatar jningwei avatar ps48 avatar maxfrei750 avatar viredery avatar yanfengliu avatar cclauss avatar concerttttt avatar jmtatsch avatar bill2239 avatar llltttppp avatar marluca avatar mowoe avatar np-csu avatar robaleman avatar xelmirage avatar yqnlp avatar ziyigogogo avatar veemon avatar stevenhickson avatar skt7 avatar rymalia avatar moorage avatar manasbedmutha98 avatar

Stargazers

Jonas Heinle avatar  avatar  avatar Alberto Briseno avatar Emily Melzer avatar

Watchers

James Cloos avatar

Forkers

muhammadha04

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.