POS-ISP: Pipeline Optimization at the Sequence Level for Task-aware ISP

TL;DR

POS-ISP introduces a sequence-level reinforcement learning framework for optimizing image signal processing pipelines, improving task performance and efficiency by predicting entire module sequences in a single step.

Contribution

It formulates ISP pipeline optimization as a global sequence prediction problem, overcoming limitations of previous NAS and RL methods.

Findings

POS-ISP outperforms existing methods on multiple tasks.

It reduces computational cost compared to stage-wise approaches.

Sequence-level optimization provides stable training and better performance.

Abstract

Recent work has explored optimizing image signal processing (ISP) pipelines for various tasks by composing predefined modules and adapting them to task-specific objectives. However, jointly optimizing module sequences and parameters remains challenging. Existing approaches rely on neural architecture search (NAS) or step-wise reinforcement learning (RL), but NAS suffers from a training-inference mismatch, while step-wise RL leads to unstable training and high computational overhead due to stage-wise decision-making. We propose POS-ISP, a sequence-level RL framework that formulates modular ISP optimization as a global sequence prediction problem. Our method predicts the entire module sequence and its parameters in a single forward pass and optimizes the pipeline using a terminal task reward, eliminating the need for intermediate supervision and redundant executions. Experiments across multiple downstream tasks show that POS-ISP improves task performance while reducing computational cost, highlighting sequence-level optimization as a stable and efficient paradigm for task-aware ISP. The project page is available at https://w1jyun.github.io/POS-ISP