View Article

Abstract

This paper presents the development of a Chrome extension designed to summarize video content by capturing the URL of the video being watched, sending it to a Python backend, and returning a concise summary of the video’s description. With the exponential increase in video content, users often struggle to quickly grasp the essential points of a video without watching the entire content. This tool seeks to enhance user experience by providing quick and accurate summaries across various video platforms, thus saving time and increasing productivity. The backend system employs web scraping to extract video metadata, which is then processed using advanced Natural Language Processing (NLP) algorithms to generate summaries. Our results show that the extension works universally across platforms, offering consistent and reliable performance.

Keywords

Chrome Extension, Video Summarization, Natural Language Processing, Web Scraping, User Experience, Video Content Management

Introduction

The sheer volume of online video content, from educational videos to entertainment, presents a challenge in efficiently consuming relevant information. Viewers often face difficulties in deciding whether a video is worth watching in full, particularly when descriptions are lengthy or lacking in detail. B. Problem Statement There is a need for a tool that can provide concise summaries of video content, allowing users to quickly determine the relevance of the video. Existing solutions are often platform specific or require manual intervention, limiting their utility. C. Objective The objective of this project is to develop a Chrome extension that automatically extracts the URL of a video being watched, communicates with a Python backend to fetch the video’s description, and presents a summarized version to the user. This tool is designed to work across any video platform, making it a universal solution for video content summarization.

LITERATURE SURVEY

1. Video Summarization Techniques Landscape

Video contents have dominated our fast-paced digital lives, but because of staggering amounts of content available online, it can get hard to process. Hence, with the increased interest in summarizing techniques of video content, it has allowed viewers easy digestion of large amounts of information quickly.[5] 

In this paper, we demonstrate the subtlety of Distinguishing    various   video summarization techniques, dig into the technological foundation, and survey the current systems while navigating through this space to discover potential new inventions.

2. The Video Summarization Techniques

Video summarization can be broadly categorized into two techniques: visual-based summarization and text-based summarization. However with each comes its strength and weakness within the method.

3. Visual-Based Summarization

Visual-based summarization is based on visual features existing in a video. This summarization technique focuses on key frames or shots, which represent the summation content as a whole.

* Keyframe Extraction/Shot Boundary Detection According to Money and Agius (2008), generic methods involve keyframe extraction and shot boundary detection. While these techniques are frequently employed, it makes them messy and computationally intensive, especially when it comes to algorithms that are specific to a particular domain, which makes generalization across different platforms difficult.

*Video Skimming*: This generates a summary of the video, enabling a viewer to grasp the overall idea in a short period without having to view the video at its full length. There are various visual summarization tools; however, they require substantial computational costs that can be off-putting to non-developer users.

4. Text-Based Summarization

Text-based summarization relies upon the textual information available with the video, such as transcriptions, captions, or description. Over the years, this technique has become pretty popular because of the advancements developed in the field of Natural Language Processing (NLP)

NLP Techniques: Mokhtari et al. (2018)[4][9][20] prefer to use NLP technology to convert raw transcriptions to a coherent summary.   They can therefore distill the essence of the video efficiently through various algorithms.

*Alignment with our Project Goals: The ultimate goal of this blog post is to back up a project that aims at creating shorthand video summaries, based exclusively on descriptions found on the internet.

5. Natural Language Processing's Role

Natural Language Processing is an extremely advanced field, and transformer-based models - BERT and GPT for example - dramatically revolutionized the task of summarization.

1. Video Summarization Techniques Landscape

Video contents have dominated our fast-paced digital lives, but because of staggering amounts of content available online, it can get hard to process. Hence, with the increased interest in summarizing techniques of video content, it has allowed viewers easy digestion of large amounts of information quickly. In this paper, we demonstrate the subtlety of distinguishing various video summarization techniques, dig into the technological foundation, and survey the current systems while navigating through this space to discover potential new inventions.             

Natural Language Processing's Role

Natural Language Processing is an extremely advanced field, and transformer-based models - BERT and GPT for example - dramatically revolutionised the task of summarization.

  • Extractive Summarization: Here, summary is generated by extracting key sentences of the text. The latest example is the TextRank algorithm, developed by Mihalcea and Tarau, in 2004[1],[11] which basically implements a ranking model similar to PageRank.
  • Abstractive Summarization: This is a type of summarization that generates new sentences with novel words which essentially capture the crux of the original content. Recent models such as BART and T5 (Lewis et al., 2020; Raffel et al., 2020) [2][3] have significantly improved their capabilities toward making a machine create summaries which are almost the same as the human outputting system.

Our project strives to benefit from these recent breakthroughs by leveraging extractive or abstractive summarization techniques to make video content summaries more concise and readable.

6. Web Scraping Techniques

Web scraping is the central building block of our video summarization system. It's employed in order to effectively extract video descriptions from all of these platforms.

7. Ethical and Legal Issues

As one makes webscrapes, ethical and legal lines need to be crossed by users. An example for user can be provided by Mishra et al. (2020)[7], which proves that while using a website's resources, it is also important to follow the rules about robots.txt files and API usage policies. Thus, users are recommended to be alert and not to risk their legal side.

8. Tools and Libraries

There are numerous web scraping libraries created in Python that are gaining popularity as a result of their effectiveness for dynamic acquisition of raw data from websites. According to Alkhateeb et al. (2019) [6] Beautiful Soup: It provides a simple way to navigate and extract data out of an HTML and XML document.

Scrapy: It is an advanced framework that helps the user in   building web scrapers quickly. The first challenge that we face with this project is the     development of a scraper, able to process diverse HTMl structures across different platforms.  

9. Browser Extensions for Video Content Summary

Although   there are numerous browser extensions in the text summarization task, applying such tools for video content possesses special challenges and opportunities.

10. Existing Tools

Some of these extensions revolve around the summarization of text content, such as distilling articles, and long web pages into    smaller versions. As Pal et al. (2015) state:

• TL;DR: This add-on specialized in giving concise summaries of web pages.

• SMMRY: This tool also summarizes web content quite well but does not have video support.

11. Technical Challenges

It presents the integration of building a general summarization application for video content, raising many issues. According to Garcia et al. (2017)[13], extensions must be interactively designed with numerous web platforms, especially for multimedia content. Our project will forward this by developing a compatible video summarization extension that integrates across multiple platforms.

12. Comparative Analysis

Situating our project in the given current scenario of video summarization, a comparison analysis is a great need between our approach versus other traditional techniques would be of great importance.

13. Newness of Our Approach

  • Text-based Information: Unlike visual summary tools, our strategy works entirely on the text summary. This makes our strategy less resource consumer and flexible and adaptable to various types of devices.
  • Universal Accessibility: Since our proposed system utilizes a Chrome extension, it is expected to have a universal applicability as opposed to already existing standalone applications.

Finally, this project aspires for a novel synthesis of web scraping, NLP, and browser extension technologies to eventually create a video summarization tool universally applicable. This solution is going to alleviate one common problem that the video consumer has-that is, devising an easily accessible method by which video content could be digested accordingly. Obviously, video summarization is now in an evolutionary stage because both visual methods and text-based methods compete for headway. The more technologies bloom                                 and boom, then the more opportunities exist for the seamless integration of video, text, and responsible web scraping techniques. Our project is an innovative discovery indeed as it provides a synthesized package that can be accessed as a tool, giving the user the capability to navigate through an endless sea of video content with ease. Indeed, while we continue to wonder about how the technological landscape has facilitated the consumption of content, we must promote new findings in video summarization methods toward a larger landscape of easy access and digestibility of information throughout our digital ecosystem for everyone provides a synthesized package that can be accessed as a tool, giving the user the capability to navigate through an endless sea of video content with ease.

EXISTING SYSTEM      

1. Navigating the World of Video Summarization Tools:

A Comprehensive Overview In an age where attention spans are shrinking and content abundant, video summarization tools have emerged as essential aids for content creators, educators, and everyday viewers. These tools allow users to distill lengthy videos into concise summaries, helping to save time and enhance understanding. In this article, we will explore the current landscape video summarization tools, categorizing them into visual-based and text-based systems, examining browser extensions for content summarization, and discussing the limitations of existing solutions. We’ll also highlight the importance of an innovative approach to video summarization that breaks down these existing barriers.

2. The Current Video Summarization Tools

Video summarization tools are designed to work either within specific platforms or as standalone software packages. These tools generally fall into two principal categories: visual-based summarization and text-based summarization.

3. Visual-based Summarization Tools

Visual-based summarization tools leverage video content analysis to identify key frames or highlights, allowing users to get a quick glimpse of the content without watching the entire video. Here are a few prominent examples:

4. Microsoft Video Indexer

Overview: Part of Azure Cognitive Services, Microsoft Video Indexer takes video analysis to the next level. It transcribes dialogues, extracts key phrases, and highlights essential moments in videos, using advanced machine learning algorithms.

Key Features:

  • Identifies key scenes based on visual and audio cues.
  • Generates short summaries, providing users with a quick overview.
  • Excellent for businesses needing video analytics for their marketing strategies.

5. SummarizeBot

Overview: This tool is powered by AI and is capable of generating text summaries from various content types, including videos.

Limitation: It lacks integration with browsers and is not universally applicable across different video platforms, making it less user-friendly for those seeking quick access.

6. Text-based Summarization Tools

Text-based summarization tools rely on textual information like transcripts, subtitles, or descriptions linked to the video. These tools are particularly useful when the visual content may not be analyzed effectively. Here are some notable examples:

7. TL;DR for YouTube

Overview: This browser extension provides a succinct summary of the description given in YouTube videos.

Key Feature: While incredibly handy for YouTube users, its functionality is limited to that platform and does not extend to others, such as Vimeo or educational platforms.

8. Glasp

Overview: A Chrome extension designed to facilitate highlighting and summarizing YouTube video content.

Limitation: The need for manual input reduces efficiency, and its functionality is restricted to specific platforms.

9. Browser Extensions for Content Summarization

Apart from tools focused on video, several extensions have emerged to summarize web content, primarily articles and textual information:

10. SMMRY

Overview: This tool specializes in transforming lengthy articles and web pages into concise summaries.

Limitation: Although effective for text-based content, SMMRY does not cater to video content, making it a standalone solution for article summarization.

TL;DR Plugin

Overview: Similar to SMMRY, this Chrome extension condenses web pages by extracting essential keywords and summarizing the content.

Limitation: Like SMMRY, it is exclusively text-focused, leaving a significant gap in handling video content.            

Limitations of Existing Systems

Despite their utility, the current video summarization tools exhibit a number of limitations which a "Universal Video Summarizer Extension" aims to address:

1. Platform Dependency

Many tools are designed for specific platforms, such as YouTube, which can limit their usability for users who frequently travel across video-sharing platforms like Vimeo or educational sites.

2. Lack of Integration

Most summarization solutions function as standalone applications or require manual input, creating friction for users wanting quick and seamless access. Users often seek tools that integrate smoothly into their browsing experience without unnecessary steps.

3. Focus on Visual Summarization

Many video summarization tools emphasize visual content analysis, which can be resource-intensive. In contexts where textual summaries are essential, these tools fall short.

CONCLUSION

As video continues to dominate online content, the need for effective summarization tools becomes increasingly critical. While current options provide valuable services, their limitations highlight the necessity for more universal and integrated solutions. The proposed "Universal Video Summarizer Extension" seeks to nullify platform dependencies, enhance user convenience, and balance the focus between visual and textual summarization methods. In a world where time is fleeting, the potential for more effective video summarization tools is immense. Whether seeking to enhance learning outcomes, drive marketing efforts, or simply save time, embracing innovative summarization solutions will undoubtedly transform how we consume video content. Consider what you need most in a summarization tool—would you prioritize integration across platforms, ease of use, or perhaps the ability to provide both visual and text-based insights? These questions can guide you toward the right solution tailored to your needs.

Explore the tools available today and stay ahead of the curve as technology continues to evolve. There’s no better time than now to invest in making your viewing experience smarter and more efficient.

PROPOSED SYSTEM

Universal Video Summarizer Extension: Video Insights at Your Fingertips. Video content occupies the screens of today's fast-paced digital landscape. However, the available time for us to consume all this content is a limited asset. There are so many platforms that offer videos on so many different topics that often we struggle to find the relevant information buried within an entire video. That's where the Video Summarizer Extension has come in—the user-friendly tool providing a succinct summary of the video, thus helping users decide which parts of the video they want to watch and which they do not need to. We will discuss in this article the architecture of such a system, its characteristic features, and benefits, as well as how it attempts to ease the user's interaction with any video content.

System Overview

The Video Summarizer Extension 360 will change video content consumption forever. It achieves this by providing automatic video description summaries across various sites, including YouTube, Vimeo, and Dailymotion. This is the simple way by which the user will then be able to know the essence of the video in a shorter time than if he were trying to watch the whole video.

How Do Videos Work?

  1. URL Capture: The Chrome extension smartly captures when a user is viewing a video. Once detected, the latter captures the video's URL.
  2. Backend Processing: This URL then gets forwarded to a Python-based server that processes the data, retrieves the appropriate video description, and forms a summary.
  3. Display Summary: The final step involves this summary being presented in a user-friendly popup that is available within the Chrome extension for quick access.

System Architecture

Universal Video Summarizer Extension has three major parts in its structure:

1. Chrome Extension (Frontend)

URL Detection: This extension automatically fetches the active video and sends their URLs to the backend server through API calls.

User Interface (UI): It is so developed as to be intuitive. One click on the extension's icon instantly opens up a summary for the description of the video.

2. Backend Server (Python-based)

URL Processing: It finds the hosting environment for the video and applies web scraping to get descriptions Web Scraping Module: Leverage on libraries such as Beautiful Soup or Scrapy to fetch descriptions from different hosting environments with great compatibility.

• API Design: Use a RESTful API structure that allows for better communication from the Chrome extension to the backend.

3. NLP Processing Module

  • Text Summarization: Advanced NLP processes combine the extracted video description into a coherent summary.[10][4][12]

Summarization Techniques include

  • Extractive Summary

Key sentences are extracted from the description.

  • Abstractive Summarization: Fully new sentences are generated which narrate the main message.

Handling Response: The polished summary is transmitted back to the extension for it to display immediately.

  • Technical Stack

The proposed system uses a technologically sound stack for efficiency:

  • Frontend/Chrome Extension:
  • Technologies: JavaScript, HTML, and CSS will be used to build the interactive parts of the extension
  • Backend/Python Server:
  • Frameworks: Either Flask or Django to make it as light in weight as possible and easy in integration
  • Web Scraping libraries: BeautifulSoup and Scrapy[6]
  • NLP libraries:
  • Tools: Use Hugging Face Transformers (BERT, GPT), SpaCy, and NLTK to assure users of the highest quality of text processing and summarization.[12][13]

Features and Functionality

Some of the key features of the Universal Video Summarizer Extension are as follows:

  • Cross-Platform Compatibility.
  • Users will have the same experience of using the video irrespective of the place of usage since it is done with a properly compatible extension for different video platforms.
  • Auto URL Detection
  • For automatic start of the summarization process, it starts immediately when it senses that a video will be used to save human input.
  • Real-Time Summarization.
  • It shows summaries almost immediately to make a decision on what to watch next.
  • Friendly User Interface.
  • The format of the summaries is clear with the easy usability in the popup, which is containing options to read full description or share the summary with friends.
  • User Experience.

The extension enables users to get access to video summaries quickly and easily, allowing them to make the most of their viewing time. It makes video viewing effortless by eliminating prolonged searches over clarity on contents. Users can browse video with confidence because summaries shall be readily available for them.

Evaluation and Testing

To ensure that the system meets the targeted user needs and it performs at its best, the following methods shall be adopted:

Performance Testing

Response time and quality of summarization will be measured through ROUGE and BLEU scores, among others.[1][8]

User Feedback

Systemic testing by the users will help us understand usability and the quality of the generated summaries, thus leading to improvement iteration based on real users' experience.

Compared to Existing Systems

The Universal Video Summarizer Extension has the following distinguished advantages over existing solutions:

Cross-Platform Compatibility: Unlike many currently available tools that appear to be focusing on a particular platform, this extension offers a single answer that suits both Windows and Mac.  

User Convenience: Forget about juggling multiple applications; one place to go for video summaries.

Quality Summary: Tap into leading-edge NLP technologies to promise high-quality effective summaries to make decisions on video content.

The Universal Video Summarizer Extension is a landmark shift in our experience with using video content onROline, making it much easier for users to access comprehensive summaries in just a few clicks, saving precious time and raising the pleasure of viewing multiple media platforms.

Ready to revolutionize your video consumption habits? Discover more of the Universal Video Summarizer Extension today.

Call to Action: To know more about digital content innovation, check our blog for latest information.

For reading more on summarization technologies using NLP, take a look at this [article](https://www.analyticsvidhya.com/blog/2020/08/understanding-text-summarization-in-natural-language-processing/).

       
            1.jpg
       

RELATIVE STUDY

YouTube TL;DR vs. Universal Video Summarizer:   

Comprehensive Comparison

In an where attention spans are dwindling and information overload is a common hurdle, video summarization tools emerged as lifesavers for viewers. In this blog post, we will compare two significant players in this field: the *YouTube          TL;DR extension* and the *Universal Video Summarizer Extension*. We will explore various aspects such as compatibility, summarization methods, user interaction, and more, to help you decide which tool suits your needs best.

Scope and Platform Compatibility

YouTube TL;DR Extension

The *YouTube TL;DR* is designed specifically for the YouTube platform, leveraging the platform's unique features to generate concise video summaries. Here are some of its keys value:

Platform-specific: The extension is only compatible with YouTube, limiting its accessibility for users who consume content across multiple video platforms such as Vimeo or Dailymotion.

Content sources: It summarizes videos by analyzing descriptions and user comments, but this approach has its downsides.

Universal Video Summarizer Extension On the other hand, the Video Summarizer 360 shines in its versatility:

Platform-agnostic: This extension works seamlessly across multiple video hosting sites. Whether you're watching a tutorial on Vimeo or a documentary on Dailymotion, this tool recognizes the video URL and applies its summarization process.

Wide applicability: This feature ensures a broader audience can benefit from its capabilities, making it an invaluable tool for anyone engaging with video content online. "The Universal Video Summarizer Extension is designed to adapt to the way you consume video content, ensuring you get the key takeaways regardless of where you are."

Summarization Methods

In terms of summarization techniques, the YouTube TL;DR extension employs a simpler method:

Key sentence extraction: It primarily extracts significant sentences or user comments that have received votes or are marked as important. While this can convey some key points, it might not accurately represent the video's overall message.

Universal Video Summarizer Extension

The Universal Video Summarizer utilizes advanced Natural Language Processing (NLP) techniques for a more contextually rich summarization:

Comprehensive metadata analysis: It generates summaries from  the video  description and metadata, ensuring   a more  thorough   understanding of the video’s content.

Improved accuracy*: This more sophisticated approach typically results in higher-quality summaries that capture the

essence of what the video offers.

User Interaction

YouTube TL;DR

When it comes to user interaction, the YouTube TL;DR extension presents some challenges:

Overlay display: The summary appears as an overlay on the video page. This can sometimes crowd the interface or get lost in the multitude of visual elements that characterize YouTube’s layout.

Universal Video Summarizer Extension

Conversely, the Universal Video Summarizer boasts a more streamlined user experience:

Popup interface: Summaries are displayed in a popup, allowing users to read them without disrupting their viewing experience.

Enhanced readability*: This elegant design does not interfere with video playing or browsing, making it easier for users to absorb information.

SMMRY vs. Universal Video Summarizer Extension

Content Type

The SMMRY tool primarily focuses on summarizing text content, which offers limited functionalities for video content:

Manual input required: Users can manually enter text, such as video transcripts, for summarization.

In contrast, the Universal Video Summarizer Extension is dedicated explicitly to video content:

Automated process: It analyzes video descriptions or metadata directly, offering an automated and frictionless experience.

Automation and Integration

SMMRY’s manual approach can detract from the user experience:

Chore-like process: Users need to copy and paste descriptions or transcripts to generate a summary, which can feel like a task.

With Universal Video Summarizer Extension, automation is its strong suit:

Seamless execution: The extension automatically detects when a video is being viewed and fetches its description to generate a summary, enhancing usability considerably.

• SummarizeBot vs. Universal Video Summarizer Extension

• Versatility

While SummarizeBot is an AI tool that handles various content types including text and images, it is not optimized for video:

Standalone requirement: Users must navigate away from their current viewing to access this tool, disrupting the user experience.

In contrast, the Universal Video Summarizer exists as a browser extension, directly integrating into the user's workflow.

•  User Experience

• SummarizeBot does have numerous summarization options, but there’s a catch:

• User-disjointed experience: The requirement to upload or enter content manually can be off-putting for those seeking quick video summaries.

• On-the-fly summaries: Users receive instant feedback with main points highlighted, allowing for quick and effective comprehension of the video content.

• Microsoft Video Indexer vs. Universal Video Summarizer Extension

•  Technical Complexity

The Microsoft Video Indexer is indeed a powerful tool, offering advanced capabilities through Azure Cognitive Services:

Developer-centric: It is tailored mainly for developers, requiring technical expertise for setup and integration. However, the Universal Video Summarizer Extension aims at broader usability:

User-friendly design: Its design caters to everyday users who need quick summaries without technical hurdles.

•  Target Audience

Microsoft’s offering primarily targets businesses and developers seeking sophisticated video analysis capabilities:

• Overkill for casual users: The complexity involved makes it unsuitable for casual viewers who only want simple video summaries.

In contrast, the Universal Video Summarizer Extension is perf

METHODOLOGY

A. System Architecture

1. Frontend (Chrome Extension):

URL Capture: The extension hooks into the browser to detect when a video is being played. It then captures the video's URL and sends it to the backend server via an API call.

UI: A simple and intuitive UI is provided within the

Chrome extension, where users can view the summarized content.

2. Backend (Python Server):

URL Processing: On receiving the URL, the backend identifies the video platform (for instance, YouTube, Vimeo) and uses web scraping approaches to fetch video description and other metadata.

Summarization Module: The video description is passed through NLP algorithms. Model usage could be something like BERT, GPT-3, or a custom-trained model that is dedicated towards text summarization.

API Response: The backend picks up the summarized text back to the Chrome extension, and then it renders it for the user to read

B. Technical Stack

Frontend: JavaScript, HTML, CSS for developing the Chrome extension.

Backend: Python with Flask or Django for API requests handling, BeautifulSoup or Scrapy for web scraping and NLP libraries such as Hugging Face Transformers, SpaCy, or NLTK for text processing.[6]

Deployment: The backend could be hosted on cloud platforms such as AWS or Heroku for scalability and reliability.

Implementation

A. Chrome Extension Development

1. Manifest File: Define the extension's permissions,

background scripts, and content scripts.

2. Video Page Script: Embedded on video pages to scan video URLs.

3. Pop-up Interface: Displays the summarized summary in a readable format.

B. Backend Server

1. URL Identification: Determine the video service and retrieve the related description.

2. Web Scraping: Utilize Python libraries to scrape the video description from the page's HTML.[7]

3. NLP Processing: Apply text summarization algorithms. The techniques comprise extractive summarization by selecting key sentences, abstractive summarization that generates new sentences.

C. API Design

* POST Request: This plugin sends the type of     HTTP request POST including video URL.

* Server Response: The server returns a JSON object   response that summarizes.

       
            2.jpg
       

OUTPUT

The "Universal Video Summarizer Extension" returns as its output a video content summary, which is presented directly to the user within the Chrome extension's popup. Here's a step-by-step analysis of output details at each step of the process:

1. Captured Video URL

 Describe: The video's URL that the user is currently

watching is captured by the Chrome extension.

 Format: The URL is a standard web link (e.g.,

https://www.youtube.com/watch?v=example).

 Role: This is the only URL that needs to be fed in for

activating the entire summary process.

2. Video Description Extraction

 Description: Using web scraping methods, the backend server retrieves the description or

metadata associated with the video.

Format: The description would typically be a text block, likely plain text, containing information about the video, either its content, intent, or background.

 Role: This text is the content that will be summarized to produce an output that is relatively short.

3. Product Description Summary

Nature: Natural Language Processing (NLP) module

processes video description and summary.

Format: Summary is a shorter version of the

original description, usually some sentences, giving a few key points or an overview of the video.

Purpose: The summary produced here is the core product to be vended to the user, making it easy for him/her to just glance through the video content.

4. Summary to be Displayed in Chrome Extension Summary:

The generated summary is sent back to the Chrome extension and displayed in a user-friendlypopup. Interface: The summary is displayed as plain text within the popup but formatted for easy reading. Such a popup created by the extension may also carry the view full description options, copy a summary, or share the summary.

Output: It is the final output that the user interacts with and provides the user with the basic essence of the video content right there in the browser.Key Output Characteristics

Accuracy: The summary should capture the essential

points of the video description, such that the language is brief yet informative.

Readability: The summary should be clear, precise in the choice of words, and easily readable.

Real-Time: The summary should appear in almost real-time while the user is browsing through videos, meaning it is up to date and has the right detail at the right time.

Example Output Scenarios

1. Scenario

Educational Video on YouTube

Video Describe: "It will cover the basics of

Python programming, including variables, data types, and control structures.

3. Scenario

Product Review Video on Vimeo

Video Description: "Review of the latest smartphone model, maintaining features, performance, and value for money".

Generated Summary: "Review of latest smartphone,

on features, performance, and value".

3. Scenario:

Inspiring Video Talk on Dailymotion

Video Description: "Join us for an inspiring talk about

overcoming obstacles and achieving your dreams, delivered by a leading motivational speaker."

Generated Summary:

"An inspiring speech on overcoming challenges and achieving dreams by a motivational speaker."

REFERENCE

  1. Mihalcea, R., & Tarau, P. (2004). TextRank: Bringing Order into Texts. EMNLP 2004. Retrieved from https://aclanthology.org/W04-3252/
  2. Lewis, M., Liu, Y., Goyal, N., Ghazvininejad, M., Mohamed, A., Levy, O., Stoyanov, V., & Zettlemoyer, L. (2020). BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension. Retrieved from https://arxiv.org/abs/1910.13461
  3. Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2020). Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer. Journal of Machine Learning Research, 21(140), 1-67. Retrieved from http://jmlr.org/papers/v21/20-074.html
  4. Mokhtari, M., Delcroix, M., & Hisada, K. (2018). Multi-Granular Word Embedding for Sentence Similarity in Summarization. EMNLP 2018.
  5. Money, A. G., & Agius, H. W. (2008). Video Summarization: A Conceptual Framework and Survey of the State of the Art. Journal of Visual Communication and Image Representation, 19(2), 121-143.
  6. Alkhateeb, M., Faizan, M., & Al-Dmour, M. (2019). Web Scraping in Python Using Beautiful Soup Library. In: Getting Structured Data from the Internet. Apress, Berkeley, CA. DOI: 10.1007/978-1-4842-6576-52
  7. Mishra, N., Reddy, S., & Kankal, S. (2020). Ethical Implications of Web Scraping and Video Summarization Tools in a Dynamic Web Environment.
  8. Paulus, R., Xiong, C., & Socher, R. (2018). A Deep Reinforced Model for Abstractive Summarization. ICLR 2018. Retrieved from https://openreview.net/forum?id=HkAClQgA
  9. Narayan, S., Cohen, S. B., & Lapata, M. (2018). Don’t Give Me the Details, Just the Summary! Topic-Aware Convolutional Networks for Extreme Summarization. EMNLP 2018. Retrieved from https://aclanthology.org/D18-1206/
  10. Cheng, J., & Lapata, M. (2016). Neural Summarization by Extracting Sentences and Words. ACL 2016. Retrieved from https://aclanthology.org/P16-1046/
  11. Zhang, Y., & Bender, E. M. (2020). Overview of Language Representation Learning. Annual Review of Linguistics, 6(1), 329-353. DOI: 10.1146/annurev-linguistics-011719-043344
  12. Liu, Y., & Lapata, M. (2019). Text Summarization with Pretrained Encoders. EMNLP 2019. Retrieved from https://aclanthology.org/D19-1387/
  13. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, ?., & Polosukhin, I. (2017). Attention Is All You Need. NeurIPS 2017. Retrieved from https://arxiv.org/abs/1706.03762
  14. Kim, Y., Kim, J., & Lee, Y. (2019). Deep Learning-Based Real-Time Video Summarization for Video Browsing. IEEE Access, 7, 29217-29229. DOI: 10.1109/ACCESS.2019.2901240
  15. Cao, Y., Li, Y., Liu, Y., Li, S., Liu, L., & Zha, H. (2021). Video Summarization by Learning Relationships between Motion, RGB, and Audio. CVPR 2021. Retrieved from https://openaccess.thecvf.com/content/CVPR2021/papers/Cao_Video_Summarization_by_Learning_Relationships_Between_Motion_RGB_and_Audio_CVPR_2021_paper.pdf
  16. Bigham, J. P., & Garber, A. (2014). Video Highlighting: Improving Accessibility for Deaf and Hard of Hearing People. ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2014). Retrieved from https://dl.acm.org/doi/10.1145/2661334.2661372
  17. Van Der Zant, T., & Wiering, M. (2006). Web-Page Summarization Using a Word-Weight Algorithm. International Conference on Web Information Systems Engineering (WISE 2006). Retrieved from https://link.springer.com/chapter/10.1007/11925231_5
  18. Patel, J. M. (2020). Web Scraping in Python Using Beautiful Soup Library. In: Getting Structured Data from the Internet. Apress, Berkeley, CA. DOI: 10.1007/978-1-4842-6576-52
  19. Ouyang, W., Li, H., & Wang, X. (2017). Video Summarization with Long Short-Term Memory. IEEE Transactions on Circuits and Systems for Video Technology, 27(6), 1302-1315. DOI: 10.1109/TCSVT.2016.2571560
  20. Goyal, R., Jain, R., Mittal, A., Bhattacharya, I., & Chakraborty, T. (2019). A Survey on Natural Language Video Description. ACM Computing Surveys, 52(6), 1-37. DOI: 10.1145/3354035
  21. Chen, J., Zhu, Q., & Yu, S. (2019). A Comprehensive Review of Web Scraping. International Journal of Future Generation Communication and Networking, 12(4), 12-20.

Reference

  1. Mihalcea, R., & Tarau, P. (2004). TextRank: Bringing Order into Texts. EMNLP 2004. Retrieved from https://aclanthology.org/W04-3252/
  2. Lewis, M., Liu, Y., Goyal, N., Ghazvininejad, M., Mohamed, A., Levy, O., Stoyanov, V., & Zettlemoyer, L. (2020). BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension. Retrieved from https://arxiv.org/abs/1910.13461
  3. Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2020). Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer. Journal of Machine Learning Research, 21(140), 1-67. Retrieved from http://jmlr.org/papers/v21/20-074.html
  4. Mokhtari, M., Delcroix, M., & Hisada, K. (2018). Multi-Granular Word Embedding for Sentence Similarity in Summarization. EMNLP 2018.
  5. Money, A. G., & Agius, H. W. (2008). Video Summarization: A Conceptual Framework and Survey of the State of the Art. Journal of Visual Communication and Image Representation, 19(2), 121-143.
  6. Alkhateeb, M., Faizan, M., & Al-Dmour, M. (2019). Web Scraping in Python Using Beautiful Soup Library. In: Getting Structured Data from the Internet. Apress, Berkeley, CA. DOI: 10.1007/978-1-4842-6576-52
  7. Mishra, N., Reddy, S., & Kankal, S. (2020). Ethical Implications of Web Scraping and Video Summarization Tools in a Dynamic Web Environment.
  8. Paulus, R., Xiong, C., & Socher, R. (2018). A Deep Reinforced Model for Abstractive Summarization. ICLR 2018. Retrieved from https://openreview.net/forum?id=HkAClQgA
  9. Narayan, S., Cohen, S. B., & Lapata, M. (2018). Don’t Give Me the Details, Just the Summary! Topic-Aware Convolutional Networks for Extreme Summarization. EMNLP 2018. Retrieved from https://aclanthology.org/D18-1206/
  10. Cheng, J., & Lapata, M. (2016). Neural Summarization by Extracting Sentences and Words. ACL 2016. Retrieved from https://aclanthology.org/P16-1046/
  11. Zhang, Y., & Bender, E. M. (2020). Overview of Language Representation Learning. Annual Review of Linguistics, 6(1), 329-353. DOI: 10.1146/annurev-linguistics-011719-043344
  12. Liu, Y., & Lapata, M. (2019). Text Summarization with Pretrained Encoders. EMNLP 2019. Retrieved from https://aclanthology.org/D19-1387/
  13. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, ?., & Polosukhin, I. (2017). Attention Is All You Need. NeurIPS 2017. Retrieved from https://arxiv.org/abs/1706.03762
  14. Kim, Y., Kim, J., & Lee, Y. (2019). Deep Learning-Based Real-Time Video Summarization for Video Browsing. IEEE Access, 7, 29217-29229. DOI: 10.1109/ACCESS.2019.2901240
  15. Cao, Y., Li, Y., Liu, Y., Li, S., Liu, L., & Zha, H. (2021). Video Summarization by Learning Relationships between Motion, RGB, and Audio. CVPR 2021. Retrieved from https://openaccess.thecvf.com/content/CVPR2021/papers/Cao_Video_Summarization_by_Learning_Relationships_Between_Motion_RGB_and_Audio_CVPR_2021_paper.pdf
  16. Bigham, J. P., & Garber, A. (2014). Video Highlighting: Improving Accessibility for Deaf and Hard of Hearing People. ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2014). Retrieved from https://dl.acm.org/doi/10.1145/2661334.2661372
  17. Van Der Zant, T., & Wiering, M. (2006). Web-Page Summarization Using a Word-Weight Algorithm. International Conference on Web Information Systems Engineering (WISE 2006). Retrieved from https://link.springer.com/chapter/10.1007/11925231_5
  18. Patel, J. M. (2020). Web Scraping in Python Using Beautiful Soup Library. In: Getting Structured Data from the Internet. Apress, Berkeley, CA. DOI: 10.1007/978-1-4842-6576-52
  19. Ouyang, W., Li, H., & Wang, X. (2017). Video Summarization with Long Short-Term Memory. IEEE Transactions on Circuits and Systems for Video Technology, 27(6), 1302-1315. DOI: 10.1109/TCSVT.2016.2571560
  20. Goyal, R., Jain, R., Mittal, A., Bhattacharya, I., & Chakraborty, T. (2019). A Survey on Natural Language Video Description. ACM Computing Surveys, 52(6), 1-37. DOI: 10.1145/3354035
  21. Chen, J., Zhu, Q., & Yu, S. (2019). A Comprehensive Review of Web Scraping. International Journal of Future Generation Communication and Networking, 12(4), 12-20.

Photo
Shraddha Kankal
Corresponding author

G.H Raisoni college of engineering and management pune

Photo
Shravani yenge
Co-author

G.H Raisoni college of engineering and management Pune

Photo
Sanskruti tehare
Co-author

G.H Raisoni college of engineering and management pune

Photo
Nilamadhab Mishra
Co-author

G.H Raisoni college of engineering and management pune

Photo
Gayatri Deshmukh
Co-author

G.H Raisoni college of engineering and management pune

Gayatri Deshmukh, Nilamadhab Mishra, Sanskruti Tehare, Shravani Reddy, Shraddha Kankal, Summarizer extension 360, Int. J. Sci. R. Tech., 2024, 1 (3), 84-93. https://doi.org/10.5281/zenodo.13987111

More related articles
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sank...
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sank...
A Review on Green Tea (camellia sinensis)...
Swapnil Wadkar, Tejaswini Gurud, Sneha Kanse, Sagar Kale, Akash B...
AI-Driven Facial Recognition-Based Photo Retrieval System for Event Management...
Vishwanath M. S., Saumya G., Ravina G., D. Linett Sophia, ...
A Review on Green Tea (camellia sinensis)...
Swapnil Wadkar, Tejaswini Gurud, Sneha Kanse, Sagar Kale, Akash Balid, Darshan Wagh, Pragati Padole,...
A Review on Green Tea (camellia sinensis)...
Swapnil Wadkar, Tejaswini Gurud, Sneha Kanse, Sagar Kale, Akash Balid, Darshan Wagh, Pragati Padole,...
Related Articles
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sanket Fulari, Naman Gandhi, Sagar Kale...
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sanket Fulari, Naman Gandhi, Sagar Kale...
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sanket Fulari, Naman Gandhi, Sagar Kale...
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sanket Fulari, Naman Gandhi, Sagar Kale...
More related articles
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sanket Fulari, Naman Gandhi, Sagar Kale...
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sanket Fulari, Naman Gandhi, Sagar Kale...
A Review on Green Tea (camellia sinensis)...
Swapnil Wadkar, Tejaswini Gurud, Sneha Kanse, Sagar Kale, Akash Balid, Darshan Wagh, Pragati Padole,...
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sanket Fulari, Naman Gandhi, Sagar Kale...
A Review on Vitiligo A Chronic Skin Disorder...
Abhishek Bhosale, Sneha Kanase, Swapnil Wadkar, Akash Balid, Sanket Fulari, Naman Gandhi, Sagar Kale...
A Review on Green Tea (camellia sinensis)...
Swapnil Wadkar, Tejaswini Gurud, Sneha Kanse, Sagar Kale, Akash Balid, Darshan Wagh, Pragati Padole,...