Pith. sign in

REVIEW

Towards RAW Object Detection in Diverse Conditions

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2411.15678 v1 pith:HPW4T26G submitted 2024-11-24 cs.CV

Towards RAW Object Detection in Diverse Conditions

classification cs.CV
keywords detectionsrgbconditionsobjectpre-trainingaodrawdomainunder
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Existing object detection methods often consider sRGB input, which was compressed from RAW data using ISP originally designed for visualization. However, such compression might lose crucial information for detection, especially under complex light and weather conditions. We introduce the AODRaw dataset, which offers 7,785 high-resolution real RAW images with 135,601 annotated instances spanning 62 categories, capturing a broad range of indoor and outdoor scenes under 9 distinct light and weather conditions. Based on AODRaw that supports RAW and sRGB object detection, we provide a comprehensive benchmark for evaluating current detection methods. We find that sRGB pre-training constrains the potential of RAW object detection due to the domain gap between sRGB and RAW, prompting us to directly pre-train on the RAW domain. However, it is harder for RAW pre-training to learn rich representations than sRGB pre-training due to the camera noise. To assist RAW pre-training, we distill the knowledge from an off-the-shelf model pre-trained on the sRGB domain. As a result, we achieve substantial improvements under diverse and adverse conditions without relying on extra pre-processing modules. Code and dataset are available at https://github.com/lzyhha/AODRaw.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.