CBL - Campus del Baix Llobregat

Projecte llegit

Títol: Análisis de vídeo con aplicaciones en inclusión y diversidad empresarial

Estudiants que han llegit aquest projecte:

Director/a: SENAYA, KOFI DZIFA

Departament: OE

Títol: Análisis de vídeo con aplicaciones en inclusión y diversidad empresarial

Data inici oferta: 10-03-2025     Data finalització oferta: 10-11-2025


Estudis d'assignació del projecte:
    GR ENG SIS TELECOMUN
Tipus: Individual
 
Lloc de realització: EETAC
 
Paraules clau:
Transcripción
 
Descripció del contingut i pla d'activitats:
Las empresas actuales manejan grandes volúmenes de contenido audiovisual en sus procesos de selección de personal y formación interna. Sin embargo, estos procesos pueden verse afectados por sesgos inconscientes y barreras de accesibilidad. Con el objetivo de abordar estos desafíos, se abre la convocatoria para un Trabajo de Fin de Grado (TFG) centrado en el desarrollo de una aplicación en Python que permita el análisis automatizado de archivos de vídeo, proporcionando herramientas para la toma de decisiones objetivas y la mejora de la accesibilidad en el entorno empresarial.
 
Overview (resum en anglès):
The need identified in this project is reflected in the optimization of transcription and meeting summary processes. To effectively document meetings in business environments, it is essential to have efficient tools that automate and simplify these tasks, especially for startups and small businesses with limited resources.
With advancements in artificial intelligence technologies, multiple solutions have been developed for large corporations, but few are designed to meet the needs and budgets of small enterprises. This project proposes an innovative and accessible solution capable of automating transcription, summary generation, and meeting translation, thereby facilitating information management in smaller business environments.
The developed tool integrates various advanced technologies: OpenAI's Whisper for automatic transcription, Transformer-based summarization models from Hugging Face, and Helsinki-NLP models for automatic translation. Additionally, a graphical interface has been designed using Tkinter to ensure simple and intuitive interaction with the system.
During the implementation process, priority was given to data processing efficiency, transcription accuracy, and multilingual accessibility. The result is a local tool that enhances internal organization and productivity for small businesses. However, a detailed implementation plan and processes for launching the application to the market are still pending.
Future system improvements include integration with videoconferencing platforms like Zoom or Microsoft Teams and optimization for handling large data volumes and high-background-noise environments. These actions are part of the project's roadmap, which involves optimization phases, pilot testing, and the development of support materials to ensure a robust and adaptable tool for business needs.
Ultimately, this work represents a significant personal contribution to the field of process automation for small businesses. The developed tool aims to facilitate access to advanced technologies through an efficient, flexible, and accessible solution designed to optimize information management and processing in business meetings.

© CBLTIC Campus del Baix Llobregat - UPC