# Comprehensive Generative Replay for Task-Incremental Segmentation with Concurrent Appearance and Semantic Forgetting

Source: [Comprehensive Generative Replay for Task-Incremental Segmentation with Concurrent Appearance and Semantic Forgetting](<https://www.sota2.com/research/paper/68fa73db1e209131cea5e863>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-02T06:40:10.958Z

Authors: Wei Li, Jingyang Zhang, Pheng-Ann Heng, Lixu Gu

Published: 2024-06-28T10:05:58.000Z

## Abstract

Generalist segmentation models are increasingly favored for diverse tasks involving various objects from different image sources. Task-Incremental Learning (TIL) offers a privacy-preserving training paradigm using tasks arriving sequentially, instead of gathering them due to strict data sharing policies. However, the task evolution can span a wide scope that involves shifts in both image appearance and segmentation semantics with intricate correlation, causing concurrent appearance and semantic forgetting. To solve this issue, we propose a Comprehensive Generative Replay (CGR) framework that restores appearance and semantic knowledge by synthesizing image-mask pairs to mimic past task data, which focuses on two aspects: modeling image-mask correspondence and promoting scalability for diverse tasks. Specifically, we introduce a novel Bayesian Joint Diffusion (BJD) model for high-quality synthesis of image-mask pairs with their correspondence explicitly preserved by conditional denoising. Furthermore, we develop a Task-Oriented Adapter (TOA) that recalibrates prompt embeddings to modulate the diffusion model, making the data synthesis compatible with different tasks. Experiments on incremental tasks (cardiac, fundus and prostate segmentation) show its clear advantage for alleviating concurrent appearance and semantic forgetting. Code is available at https://github.com/jingyzhang/CGR.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Medical Image Segmentation on Cardiac, Fundus, and Prostate C → F → P](<https://www.sota2.com/research/sota/medical-image-segmentation-on-cardiac-fundus-and-prostate-c-f-p.md>): Medical Image Segmentation; Cardiac, Fundus, and Prostate C → F → P; DSC Cardiac: 87.48.

- [Medical Image Segmentation on Prostate, Fundus, and Cardiac Order: P -\> F -\> C](<https://www.sota2.com/research/sota/medical-image-segmentation-on-prostate-fundus-and-cardiac-order-p-f-c.md>): Medical Image Segmentation; Prostate, Fundus, and Cardiac Order: P → F → C; DSC (Prostate): 88.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2406.19796>)

[PDF](<https://arxiv.org/pdf/2406.19796.pdf>)

[Implementation](<https://github.com/jingyzhang/CGR>)
