MishMash hackathon is India’s largest online diversity hackathon. The focus will be to give you, regardless of your background, gender, sexual orientation, ethnicity, age, skill sets and viewpoints, an opportunity to showcase your talent. The Hackathon is Live from 6:00 PM, 23rd March to 11:55 PM, 1st April, 2020
2
stars
12
forks
source link
Innovation - Automatic Highlights Generation from a Full Match Sports Video- Deep Tech or Machine Learning #187
Generating highlights manually from a full match sports video is a labor-intensive task that requires some degree of specialization. It is a time-consuming task as well and unless you work for a video company that does this job day in and day out, you need to find a different approach. It is also memory-hogging to store a full match video. So, extracting highlights automatically from a full match video saves a lot of time for the creator as well as the user.
I have built 3 approaches to solve the problem -
Simple Speech Analysis
Scorecard Detection
Overs Detection
🔦 Any other specific thing you want to highlight?
Not every problem requires Deep Learning and Machine Learning. Most of the problems can be solved by a thorough understanding of Domain and Data.
The complex problem of highlights generation is solved without using Machine Learning and Deep Learning.
Innovation - Automatic Highlights Generation from a Full Match Sports Video- Deep Tech or Machine Learning
Project information:
Theme: Deep Tech or Machine Learning
Project Name: Automatic Highlights Generation from a Full Match Sports Video
Short Project Description: To build a system that automatically generates highlights from a full match sports video
Team Name: Innovation
Team Members: Aravind (aravindpai)
Demo & Repository Links: Included in the presentation
Presentation Link: https://docs.google.com/presentation/d/1WFWIect27yH9kwfSgHiPy7Kp7woia9DhZt7tKoOvkYs/edit?usp=sharing
Your Pitch:
Generating highlights manually from a full match sports video is a labor-intensive task that requires some degree of specialization. It is a time-consuming task as well and unless you work for a video company that does this job day in and day out, you need to find a different approach. It is also memory-hogging to store a full match video. So, extracting highlights automatically from a full match video saves a lot of time for the creator as well as the user.
I have built 3 approaches to solve the problem -
🔦 Any other specific thing you want to highlight?
Not every problem requires Deep Learning and Machine Learning. Most of the problems can be solved by a thorough understanding of Domain and Data.
The complex problem of highlights generation is solved without using Machine Learning and Deep Learning.