kimeshan / mf-proto

πŸ‰ Sports video analysis and social platform

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Matchflow Prototype πŸ‰

GitHub license PRs Welcome

Sports video ⏩ analysis platform

Getting Started

You will need to install:

  1. Install the above requirements. Confirm that Node.js was installed correctly by typing the following in Windows Command prompt/Powershell:
node --version

A version should be displayed.

  1. Clone this repo using 'git clone https://github.com/kimeshan/mf-proto.git' in your chosen directory. e.g C:/mf-proto.

  2. Run MongoDB server, from Command Prompt on Windows, navigate to the directory where you installed MongoDB which contains the mongod.exe file.

  • Now, specify the data directory using the command:
    .\mongod.exe --dbpath "C:/mf-proto"
    
    This should get the MongoDB server up and listening to the default port. Do not close this command prompt window.
  1. Open a second Powershell/Command Prompt on Windows
  • Set the current directory to the chosen directory in step 2 (i.e. our Git repo). e.g cd "C:/mf-proto"
  • Run the command "node server.js".
  • "Magic happens at port 3000" should be displayed. Your firewall may pop up with a notification - if so proceed to allow Node.js to use port 3000.
  • "Connection successful to MongoDB" should also be displayed indicated that the Mongo server has been connected to.
  1. Open your web browser (preferably Chrome), and go to 'localhost:3000'. This should start the mf-proto web application! That's it.

  2. Videos uploaded will automatically upload to: your-chosen-directory-in-2/public/Video Uploads

  3. Converted videos can be located in your-chosen-directory-in-2

Technical Overview

mf-proto is a sports match and player performance video analytics platform. It's a web application built on the MEAN stack. Node.js and the Express web framework is used on the backend while AngularJS, HTML 5 and CSS are used on the front-end. MongoDB will be used as the database. mf-proto can be used offline and online in the browser while a sync process will ensure the online server is updated.

The completed web application will consist of three main components:

  1. Project, Collections & Events Management (CRUD)
  • Project Manager: CRUD for projects. Each project is stored in 'projects' table consisting of the following columns:
    • "name": {"type":"String","unique":true}
    • "date_played":{"type":"Date"}
    • "local_team":"String"
    • "opposition_team":"String"
    • "season":"String"
    • "competition":"String"
    • "local_team_score":"Number"
    • "opposition_team_score":"Number"
    • "collection_name": "String"
    • "video": "String"
  • Collections Manager: CRUD for collections of events. Collections are stored in 'collections' table which has just two columns:
    • "name":{"type":"String","unique":true}
    • "date_created": {"type":"Date","default":Date.now}
  • Events Management: CRUD for events under a collection. Events are linked to a collection by the "collection_name" column which matches the "name" column in 'collections' table:
    • "collection_name":"String"
    • "event_name":"String"
    • "lead_time":"Number"
    • "lag_time":"Number"
  1. Video Manager
  • Uploader - Performs video upload (offline and online). Offline uploads will be stored in the mf-proto application folder (which will ultimately be in C:/Program Files when the completed product is distributed). When an internet connection is available uploaded videos will be syncrhonized with a cloud-based server.
  • Converter - Converts original uploaded video using fluent-ffmpeg module. The goal here is to compress large video files (such as those from HD camcorders) to obtain a smaller file size while maintaining an acceptable viewing quality for later analysis in the Editor.
  1. Video Analysis Mode - User can tag segments of the video with these events. Events are defined by the user in a Collection visa the Collections & Events Manager (CRUD), for example an event called "Penalty Team Blue" might be created with a lead time of 10sec and lag time of 20 sec. When a time point (say 00:05:02) in the video is tagged with this event, the tag stored will define start time (lead time) as 10 seconds before (00:05:52) and end time (lag time) as 20 seconds after (00:05:22). A 'tags' table in the database will store this information.
  • Tags table columns:
    • tag_id
    • project_name (string) - Linked to 'projects' table.
    • collection_name (string) - Linked to 'collections' table.
    • event_name (string) - Linked to 'events' table.
    • start_time (timestamp of starting point in video of the play being tagged)
    • end_time (timestamp of ending point in video of the play being tagged)
    • video_name (name of video as stored in offline folder/cloud server)
    • tag_duration (difference between start_time and end_time)

Completed Components

  • Projects Manager
    • Creating, listing, deletion of projects
    • Form validation for required fields.
    • Seamless modals used, no page loading.
    • Error handling
    • Confirmation prompt before deleting.
  • Collections Manager
    • Creating, listing, deletion and updating of collections (of events).
    • Same features as projects manager ((modals, error handling, delete prompt).
  • Events Manager
    • Once a collection is selected, the events manager can be used to create, list, delete and update events inside the collection.
    • Same features as project manager (modals, error handling, delete prompt).
  • Video Manager
    • Video upload: file is copied to mf-proto application folder inside "Video Uploads" folder.
    • Video renaming according to user input via form. Validation: Upload button enabled only when a name and valid video file has been selected.
    • Video listing and deletion.
    • In-browser video player displays video for mp4, ogg and webm video files.
    • Video converter: Encoding and Compression, compresses large-sized, HD videos to optimal SD.
      • Extracts resolution, aspect ratio, frame rate, format and codec of original video.
      • Uses extracted information to calculate optimal settings* for encoding to achieve best compression and quality:
        • Original aspect ratio is maintained
        • Optimal format chosen is mp4 using H.264 codec
        • Optimal resolution for 16:9 aspect ratio videos chosen: 640x480 (SD)
        • Optimal resolution for 4:3 aspect ratio videos chosen: 640x360 (SD)
        • Videos with width lower than 640 pixels maintain original resolution to avoid distortion.
        • Optimal frame rate is 30 fps or less. Videos with higher original frame rates are converted to 30 fps or half of original frame rate. e.g. 70 fps -> 35 fps.
    • Note: Optimal settings algorithm developed based on information from [Youtube's advanced encoding settings] (https://support.google.com/youtube/answer/1722171?hl=en) and Vimeo's compression guidelines.

Future developments

  • Project manager: Update feature to edit project fields.
  • Video analysis mode: Allow tagging of time points in the video with the events that were defined by the user.
  • Video player/browser: Integrating a feature-rich video player such as Video.js to enable advance video playing and timeline features in the browser.
  • Progress bar: For video upload.
  • Sync process to synchronous videos on the user's machine with online, cloud-based server. This will allow the user to access videos from any location by logging in through the mf-proto web application on a web broswer.

About

πŸ‰ Sports video analysis and social platform


Languages

Language:JavaScript 42.5%Language:CSS 34.5%Language:HTML 23.0%