https://github.com/miladsade96/danaxa_challenge

Danaxa Technical Interview Challenge
https://github.com/miladsade96/danaxa_challenge

Last synced: 8 months ago
JSON representation

Danaxa Technical Interview Challenge

• Host: GitHub
• URL: https://github.com/miladsade96/danaxa_challenge
• Owner: miladsade96
• License: mit
• Created: 2021-07-17T16:12:17.000Z (about 4 years ago)
• Default Branch: main
• Last Pushed: 2021-07-17T18:58:20.000Z (about 4 years ago)
• Last Synced: 2025-01-04T00:27:43.222Z (9 months ago)
• Language: Python
• Size: 22.5 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# danaxa_challenge

Danaxa Technical Interview Challenge

## Challenge Description

Suppose we are developing a new labeling tool to annotate masks in a video.

Labeling all the frames of a video with a great accuracy takes a lot of time and

cost. In order to make annotation process faster, we need to use semi-

automated or automated labelling methods. Therefore, we need to implement

a method to annotate an object in a few frames and the tool keep detecting

that object in next frames. Ultimately, we want the tool to annotate all video

frames itself after annotating a few frames.

## Answering Questions

* Review previous work flow on this research topic and report it to us:

    * One shot learning 

    * Few shot learning

    * Siamese neural network

    * video object segmentation  

* Find recent works and publications which address this problem:

    * [Boundary-preserving Mask R-CNN](https://arxiv.org/abs/2007.08921)

    * [CANet: Class-Agnostic Segmentation Networks with Iterative Refinement and Attentive Few-Shot Learning](https://arxiv.org/abs/1903.02351)

    * [Learning What to Learn for Video Object Segmentation](https://arxiv.org/abs/2003.11540)

    * [Meta-DETR: Few-Shot Object Detection via Unified Image-Level Meta-Learning](https://arxiv.org/abs/2103.11731)

    * [Video Object Segmentation with Adaptive Feature Bank and Uncertain-Region Refinement](https://arxiv.org/abs/2010.07958)

    * [One-Shot Object Detection with Co-Attention and Co-Excitation](https://arxiv.org/abs/1911.12529)

    * [One-Shot Video Object Segmentation](https://arxiv.org/abs/1611.05198)

    * [Show&Tell: A Semi-Automated Image Annotation System](https://www.researchgate.net/publication/3338590_ShowTell_A_Semi-Automated_Image_Annotation_System)

    

* Find open source repositories in this regard:

    * [cvat](https://github.com/openvinotoolkit/cvat)

    * [Siamese-Networks-for-One-Shot-Learning](https://github.com/tensorfreitas/Siamese-Networks-for-One-Shot-Learning)

    * [keras-oneshot](https://github.com/sorenbouma/keras-oneshot)

    * [DPGN](https://github.com/megvii-research/DPGN)

    * [SSTVOS](https://github.com/dukebw/SSTVOS)

    * [pytracking](https://github.com/visionml/pytracking)

    

## Project Structure and Contents

images: contains a test image  

videos: contains a test video  

mrcnn: contains mask-rcnn model files  

weights: contains .h5 file (mask-rcnn model weights)  

arvf.py: contains a class in order to read video frames asynchronously  

main.py: contains loading, configuring and testing mask-rcnn model on test video  

mask_label_image.py: contains testing mask and label functionality on test image