​​​​​​​

1. Title

□ Creating AI-Pipeline applications using NNStreamer

2. Overview

□ NNStreamer is a collection of Gstreamer plugins thatenables you to easily implement and efficiently execute a data stream-based AIsystem, which consists of various neural networks, input / output terminals,and various data processors. NNStreamer developers are building it further sothat more users can easily use and apply it to more devices.

□ Anyone can also participatein the NNStreamer project in order to experience Neural Network-relateddevelopment and participate in the NNStreamer development throughcontributions.

   If you upstream to the nnstreamer.git aftercarrying out the following mission, we will provide you with a token after theevaluation of the maintainer and the voting of participating developers. Welook forward to your active participation in the development of the NNStreameropen source.

□ Mission Contents

  - Creating GUI NNStreamer AI PipelineApplications (Android/ Tizen/ Ubuntu/ macOS environments)

 When +alpha+ supportsmultiple models and suggests practical values,  or when +alpha+ hasmulti-modal interaction that has practical potential,

 it would be good ifit is suitable for the +alpha+ main webpage.

 If it is anapplication for + alpha + Android, please upload it to Play Store.

For +alpha+ Tizen targeted for use in wearable devices, usenew devices such as Galaxy Watch 3 or devices with Tizen 5.5 and above, andupload the app to the Galaxy Store.

In the development of + alpha + Application, we welcome bugmodifications / function implementation and improvement in the nnstreamer.git.​

- "+alpha+" is recommendation content thatrequires consideration.  

- Applications that have practical usage scenarios or showgood visualization as specific AI technology demonstrations are recommended.

□ Additional Requests

The application should be opened with the Apache 2.0license and merged into the NNStreamer-example.git.

Other developers must be able to build the given sourcecode and execute the results.

□ How to Participate

This mission is targeted toKorean developers. Click on the following onoff mix link to apply forparticipation and proceed with development in the NNStreamer github.

  onoffmix Link

 The application period is 9/7 ~ 10 / 14, and when you applyfor participation, you must submit the project name NNStreamer and the GitHubID that you will be using to the on-off mix application page.

□ Maintainer Review and Merge Criteria

Upstream the applications togithub.com/nnstreamer/nnstreamer-example.

All CI tests mustbe passed. (Code style, static test, build / run test)

The approval ofmore than one reviewer needs to be obtained.

Include a README.md that contains an introduction and somescreenshots of each application. (reduce the size as much as possible whenusing webp / gif).

Instead of making a large PR at once, add commits/ PRs oneby one in incremental code units. Too much commits, PRs, unreadable codes, andcommits without proper description and title are likely to be skipped forreview.

The Developer Certificate of Origin (DCO) required byGithub is also needed (accounts without DCO are automatically rejected byGithub during the PR process), and having a DCO means authenticates the factthat the developer only submits code that he / she wrote. In other words,external code should not be copy-pasted. External code required as a modulealready exist independently and should be used as a reference (e.g., a sharedlibrary). It is recommended that you consult the reviewers in advance regardingthe external code that needs to be included.

□ Recommendations for projectexecution

It is recommended that you upstream the modifications toNNStreamer itself (including bug fixes) directly togithub.com/nnstreamer/nnstreamer.

For various issues and inquiries, using the issues sectionor mailing list in github.com/nnstreamer/nnstreamer would get you a fasterresponse. If you use the issue section of nnstreamer-example, it might takesome time.

We also welcomecontributions to nnstreamer and the nntrainer.

In terms of softwareengineering, a preferred form of code is also recommended. In particular,nnstreamer and nntrainer consider the module structure, various architectureprinciples, and static analysis results from Convention important.

3. Schedule

□ Period: 9/7 ~ 10/21

□ Detailed Schedule

- Development theme opened: 9/7

- Recruitment of participants (onoff mix): 9/7 ~10/14

- Development and submission: 9/7 ~ 10/14

- Maintainer Review: 10/15 ~ 10/20

 - Results Announcement : 10/21

4. Mentoring (Support)

□ Materials to help in development

   1) Getting Started 2) wiki 3) Others Document

□ Communication channels

  - Q&A Link

□ Mailing List

  - Recommended foruse after registration

5.  Results Announcement

□ Will be announced in thewebpage after the maintainer review of the code (10 / 21)

□ A token will be provided to the developers with mergedcode after the maintainer review. In addition, excellent developers may begiven opportunities to present in various open source events such as SOSCON.

□ Commemorative items areprovided: 1) JBL Bluetooth speaker, 2) SOS Program T-shirt, 3) Stationery (notepadand ballpoint pen), 4) Sticker set.

□ As a project that isconnected with the SOSCON 2020 "SOSFest" event, we can provideadditional event tokens to developers with excellent code quality.

[Maintainer review results]

 Those who're on the list,congratulations. We will contact you individually through our program agency.

  1. dreamfinder2019(Github ID)

  2. soo-ni(Github ID)

  3. park-sang-hyun(Github ID)

  4. joojihwan (Github ID)

  5. davichiar(Github ID)