Skip to content

ynklab/Stop-Position-Ranker

Repository files navigation

Stop-Position-Ranker

Overview

This project aims to create a novel architecture and pipeline leveraging existing models to predict stop positions using scene graphs. Our approach focuses on enhancing decision-making around stop positions by introducing the ability to rank multiple candidates and provide justifications for not selecting certain positions.

Data Collection

To address the limitations of the current dataset, we are planning to augment the data with:

  • Multiple Stop Position Candidates: Collect data that includes various stop position options for each scene
  • Ranking Annotations: Add annotations that rank the candidates based on appropriateness for stopping
  • Non-Selection Reasons: Annotate why certain stop positions are not suitable

This enhanced dataset will enable our model to learn a more nuanced decision-making process, improving stop position prediction performance.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •