Dataset release for the BMVC 2021 Paper "Few-Shot Domain Adaptation for Low Light RAW Image Enhancement"

Abstract: Enhancing practical low light raw images is a difficult task due to severe noise and color distortions from short exposure time and limited illumination. Despite the success of existing Convolutional Neural Network (CNN) based methods, their performance is not adaptable to different camera domains. In addition, such methods also require large datasets with short-exposure and corresponding long-exposure ground truth raw images for each camera domain, which is tedious to compile. To address this issue, we present a novel few-shot domain adaptation method to utilize the existing source camera labeled data with few labeled samples from the target camera to improve the target domain’s enhancement quality in extreme low-light imaging. Our experiments show that only ten or fewer labeled samples from the target camera domain are sufficient to achieve similar or better enhancement performance than training a model with a large labeled target camera dataset. To support research in this direction, we also present a new low-light raw image dataset captured with a Nikon camera, comprising short-exposure and their corresponding long-exposure ground truth images. The code is available at https://val.cds.iisc.ac.in/HDR/BMVC21/index.html.

This dataset is designed for training and evaluating object detection models focused on identifying various types of litter in real-world environments.

Dataset Overview:

Total Images: 1,499

Annotations: Each image is annotated with bounding boxes corresponding to different litter categories.

Classes: 59 distinct classes representing various waste items.

Dataset Split:

Training Set: 1,049 images (70%)

Validation Set: 299 images (20%)

Test Set: 151 images (10%)

Preprocessing:

Auto-Orient: Applied to ensure consistent image orientation.

Class Modification: 59 classes remapped; none dropped.

Augmentations: No augmentations were applied in this version.

This dataset is suitable for developing and testing object detection models aimed at recognizing and classifying litter in various settings, such as urban streets, parks, and natural environments. It can be instrumental in applications related to environmental monitoring, waste management, and sustainability initiatives.

We collected a new low-light raw denoising (LRD) dataset for training and benchmarking. In contrast to the SID dataset, which sets a fixed exposure time to capture long and short exposure images, we captured long and short exposure images based on the exposure value (EV). Motivated by multi-exposure image fusion, the exposure value for long exposure images was set to 0, and the exposure value for short exposure was set to the commonly used parameters -1, -2, and -3. The dataset is designed for application to low-light raw image denoising and low-light raw image synthesis. The dataset contains both indoor and outdoor scenes. For each scene instance, we first captured a long-exposure image at ISO 100 to get a noise-free reference image. Then we captured multiple short-exposure images using different ISO levels and EVs, with a 1-2 second interval between subsequent images to wait for the sensor to cool down, thus avoiding unexpected noise introduced by sensor heating.

MUAS Markers Raw Image

## Overview

MUAS Markers Raw Image is a dataset for object detection tasks - it contains Small Object Marker annotations for 252 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

gt_RGB

View full-resolution images downlinked from the Mars Science Laboratory, sorted by Sol and by camera type.

MDP Raw Images

## Overview

MDP Raw Images is a dataset for object detection tasks - it contains Alphabets Numbers Shapes annotations for 7,244 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

A dataset for digital image forensics, containing 224 images.

The dataset contains the processed snap shots from several news channels with commercial L bands and I bands that comes during the broadcasting of news and this dataset particularly focuses on the TV commercials and advertisements in L shped form. This dataset is useful for further analysis on this domain

The goal of this project is to present two new datasets that seek to expand the capability of the Learning to See in the Dark Low-light enhancement CNN for the Canon 6D DSLR, and explore how the network performs when modified in various ways, both pruning it and making it deeper.

The original paper Learning to See in the Dark was published in CVPR 2018, by Chen Chen, Qifeng Chen, Jia Xu, and Vladlen Koltun.

Raw image files to:Figure 1. Localisation of Ago2 to the iRBC. (A) IFA of P. falciparum 3D7 blood stages. Yellow: Ago2 (EPR10411), magenta: PfTubulin. (B) IF staining of P. berghei ANKA rings and trophozoites. Yellow: Ago2 (EPR10411), magenta: PbHsp70. Nuclei were stained with Hoechst (blue). Images were taken on a Nikon spinning disc confocal microscope (100x objective). Representative images of at least 5 per condition are shown. Controls using only a secondary antibody revealed no unspecific staining. Scale bar indicates 5 µm.

Abstract ======== This data set includes the Experiment Data Record (EDR) version of all available images acquired during the Ceres Prime Mission (Approach, Transfer to Survey, Survey, HAMO, and LAMO),and extended missions 1 (XMO1, XMO2, XMO3, XMO4) and 2 (XMO6, XMO7). In addition to the imagery, ancillary information is stored within the image headers. Calibration files needed for further processing are archived as a separate data set.

View full-resolution images downlinked from the Mars Science Laboratory, sorted by Sol and by camera type.

Background Raw Images

## Overview

Background Raw Images is a dataset for object detection tasks - it contains Metal annotations for 578 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

This collection contains the raw images produced by the TAGCAMS instrument suite onboard the OSIRIS-REx spacecraft. These images were acquired for optical navigation, natural feature tracking, or sample stowage documentation.

This data set contains raw near-IR spectral image cubes acquired from November 2008 through August 2009 by the Moon Mineralogy Mapper (M3) instrument during the Chandrayaan-1 mission to the Moon.

This data set contains EDR (raw) pre-encounter and encounter images taken by the Stardust Navigation Camera of asteroid 5535 Annefrank. This is a new version of a subset of an existing data set, based on work done during the Stardust-NExT mission. The updated calibration from the Stardust-NExT mission was applied to these data in a separate data set (similar name: Level 3; RDR). Changes in the calibration of the NAVCAM instrument, between these prime mission data and those of the Stardust-NExT mission, are a possibility and are addressed in [KLAASENETAL2011B].

This data set contains raw calibration images acquired by the Deep Impact Medium Resolution Visible CCD from 04 October 2007 through 09 January 2008 for the EPOXI mission.

This collection contains the raw (processing level 0) science image data products produced by the OCAMS instrument onboard the OSIRIS-REx spacecraft.

This data set belongs to the paper "Complex pathways and memory in compressed corrugated sheets" by H. Bense and M. van Hecke, published in "Proceedings of the National Academy of Sciences of the United States of America". For more information on the methodology and techniques we refer to this article and the supplemental information belonging to the article.