Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Model Comparison for Fast Domain Adaptation in Table Service Scenario

Woo-han Yun, Minsu Jang, Jaehong Kim

TL;DR

This work tackles fast domain adaptation for automated table-service in restaurants by building a base table-information recognition model using Deformable DETR and then adapting it to local venues with limited labeled data via transfer learning. It introduces a separate table-service classifier trained on a small local dataset, and systematically evaluates feature combinations, temporal inputs, and retraining strategies, including active learning. Key findings show that freezing the backbone and blending image-based features with light attention yields strong performance in data-scarce settings, and that diversity-based active-learning sampling can outperform using the full dataset. The research demonstrates a practical pathway for deploying automated table-service systems across restaurants with minimal per-venue labeling, informing fast, scalable domain adaptation in service robotics and computer-vision for hospitality contexts.

Abstract

In restaurants, many aspects of customer service, such as greeting customers, taking orders, and processing payments, are automated. Due to the various cuisines, required services, and different standards of each restaurant, one challenging part of making the entire automated process is inspecting and providing appropriate services at the table during a meal. In this paper, we demonstrate an approach for automatically checking and providing services at the table. We initially construct a base model to recognize common information to comprehend the context of the table, such as object category, remaining food quantity, and meal progress status. After that, we add a service recognition classifier and retrain the model using a small amount of local restaurant data. We gathered data capturing the restaurant table during the meal in order to find a suitable service recognition classifier. With different inputs, combinations, time series, and data choices, we carried out a variety of tests. Through these tests, we discovered that the model with few significant data points and trainable parameters is more crucial in the case of sparse and redundant retraining data.

Model Comparison for Fast Domain Adaptation in Table Service Scenario

TL;DR

This work tackles fast domain adaptation for automated table-service in restaurants by building a base table-information recognition model using Deformable DETR and then adapting it to local venues with limited labeled data via transfer learning. It introduces a separate table-service classifier trained on a small local dataset, and systematically evaluates feature combinations, temporal inputs, and retraining strategies, including active learning. Key findings show that freezing the backbone and blending image-based features with light attention yields strong performance in data-scarce settings, and that diversity-based active-learning sampling can outperform using the full dataset. The research demonstrates a practical pathway for deploying automated table-service systems across restaurants with minimal per-venue labeling, informing fast, scalable domain adaptation in service robotics and computer-vision for hospitality contexts.

Abstract

In restaurants, many aspects of customer service, such as greeting customers, taking orders, and processing payments, are automated. Due to the various cuisines, required services, and different standards of each restaurant, one challenging part of making the entire automated process is inspecting and providing appropriate services at the table during a meal. In this paper, we demonstrate an approach for automatically checking and providing services at the table. We initially construct a base model to recognize common information to comprehend the context of the table, such as object category, remaining food quantity, and meal progress status. After that, we add a service recognition classifier and retrain the model using a small amount of local restaurant data. We gathered data capturing the restaurant table during the meal in order to find a suitable service recognition classifier. With different inputs, combinations, time series, and data choices, we carried out a variety of tests. Through these tests, we discovered that the model with few significant data points and trainable parameters is more crucial in the case of sparse and redundant retraining data.
Paper Structure (18 sections, 1 equation, 4 figures, 7 tables)

This paper contains 18 sections, 1 equation, 4 figures, 7 tables.

Table of Contents

  1. INTRODUCTION
  2. Related Work
  3. Object Detection
  4. Active Learning
  5. Approach
  6. Table Information Recognition
  7. Model
  8. Dataset and Training
  9. Table Service Suggestion Model
  10. Dataset and Training
  11. Task Definition
  12. Retraining
  13. Frozen the backbone
  14. Feature Selection
  15. Image to sequence
  16. ...and 3 more sections

Figures (4)

  • Figure 1: Table Information Recognition and Service Suggestion Model
  • Figure 2: Example images from three datasets, CloudStatus, CloudTableThings, and CloudMeals.
  • Figure 3: Example images for table service suggestion model.
  • Figure 4: Result images on four service cases. From top to bottom, each result shows the service cases of food refill, garbage collection, providing dessert, and finding lost, respectively.