# Segmentation Assisted Incremental Test Time Adaptation in an Open World

## TL;DR

This paper introduces a novel framework for incremental test time adaptation of vision language models that actively incorporates unseen classes and domains using segmentation-assisted active labeling, improving real-world adaptability.

## Contribution

It proposes SegAssist, a segmentation-assisted active labeling module that enables models to adapt to new classes and domains without retraining, a first in this context.

## Key findings

- SegAssist improves adaptation performance on benchmark datasets.
- The framework effectively handles unseen classes and distribution shifts.
- Experimental results demonstrate enhanced real-world applicability.

## Abstract

In dynamic environments, unfamiliar objects and distribution shifts are often encountered, which challenge the generalization abilities of the deployed trained models. This work addresses Incremental Test Time Adaptation of Vision Language Models, tackling scenarios where unseen classes and unseen domains continuously appear during testing. Unlike traditional Test Time Adaptation approaches, where the test stream comes only from a predefined set of classes, our framework allows models to adapt simultaneously to both covariate and label shifts, actively incorporating new classes as they emerge. Towards this goal, we establish a new benchmark for ITTA, integrating single image TTA methods for VLMs with active labeling techniques that query an oracle for samples potentially representing unseen classes during test time. We propose a segmentation assisted active labeling module, termed SegAssist, which is training free and repurposes the segmentation capabilities of VLMs to refine active sample selection, prioritizing samples likely to belong to unseen classes. Extensive experiments on several benchmark datasets demonstrate the potential of SegAssist to enhance the performance of VLMs in real world scenarios, where continuous adaptation to emerging data is essential. Project-page:https://manogna-s.github.io/segassist/

---
Source: https://tomesphere.com/paper/2508.20029