NASA Perseverance
Samples from NASA Perseverance and set of GAN generated synthetic images from Neural Mars.
태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
Image Generation (synthesis) is the task of generating new images from an existing dataset.
In this section, you can find state-of-the-art leaderboards for unconditional generation. For conditional generation, and other types of image generations, refer to the subtasks.
<span style="color:grey; opacity: 0.6">( Image credit: StyleGAN )</span>
출처에 같은 태스크 이름으로 연결된 데이터셋입니다. 동명 태스크가 포함될 수 있으므로, 비교할 때는 평가 결과의 조건도 확인하세요.
Samples from NASA Perseverance and set of GAN generated synthetic images from Neural Mars.
A new SIS benchmark designed to assess generation performance under noisy conditions, simulating human error that can occur during real-world applications. [DS] employs downsampled semantic maps that are resized by nearest-neighbor interpolation, simulating human errors e.g., jagged edges and coarse/low resolution user inputs.
A new SIS benchmark designed to assess generation performance under noisy conditions, simulating human error that can occur during real-world applications. [Edge] masks the edges of instances with an unlabeled class, imitating incomplete annotations around edges, especially between instances.
A new SIS benchmark designed to assess generation performance under noisy conditions, simulating human error that can occur during real-world applications. [Random] randomly adds an unlabeled class to the semantic maps, mimicing unintended user error and extreme random noise.
Click to add a brief description of the dataset (Markdown and LaTeX enabled). Provide: * a high-level explanation of the dataset characteristics * explain motivations and summary of its content * potential use cases of the dataset
This dataset contains prompts designed to evaluate and challenge the safety mechanisms of generative text-to-image models, with a particular focus on identifying prompts that are likely to produce images containing nudity. Introduced in the 2024 ICML paper Prompting4Debugging: Red-Teaming Text-to-Image Diffusion Models by Finding Problematic Prompts, this dataset is not specific to any single approach or model but is intended to test various mitigating measures against inappropriate content generation in models like Stable Diffusion. This dataset is only for research purposes.
This dataset builds upon the SpaGBOL dataset - a graph-based dataset covering numerous cities across the globe for the purpose of structured city-scale Cross-View Geo-Localisation (CVGL). We introduce the idea of Multi-View Geo-Localisation (MVGL) - localising in two steps within a graph. First using CVGL to retrieve the most likely primary nodes (road junctions), before performing relative pose estimation also candidate edges to increase precision. The corresponding initial network proposed reduced median error from hundreds of metres down to approximately 22m error across the entire dataset.
Pick-a-Filter is a semi-synthetic dataset constructed from Pick-a-Pic v1 to measure the capability of text-to-image models of adapting to heterogeneous preferences. We assign users from V1 randomly into two groups: those who prefer blue, cooler image tones (G1) and those who prefer red, warmer image tones (G2). After constructing this split, we apply the following logic to construct the dataset: 1. Apply “winning” and “losing” filters to appropriate images depending on label. For G1 the winning filter is blue, and for G2 the winning filter is red. 2. Randomly shortlist β % of samples to add filters. The remaining (1 − β) % of samples will remain unaltered (default images from Pick-a-Pic v1). The hyperparameter β is called the mixture ratio. 3. Randomly select 50% of above-shortlisted samples to apply a filter to only the winning image, and the remaining 50% to apply a filter to only losing image We add these sources of randomness to make learning preferences on Pick-a-Filter less prone to hacking (e.g. the model could trivially learn to predict an image with a filter as the preferred image). We provide the OpenCLIP H/14 embeddings of all versions of Pick-a-Filter used in PAL in this repository, with β varying from 0.0 (no filtering) to 1.0 (all images are filtered). For example, pick-a-filter with β = 0.4 is available in paf_0.4_mix_ratio.zip.
A synthetic dataset with 206K pressure images with 3D human poses and shapes. Source: Bodies at Rest: 3D Human Pose and Shape Estimation from a Pressure Image using Synthetic Data
Current autonomous driving algorithms heavily rely on the visible spectrum, which is prone to performance degradation in adverse conditions like fog, rain, snow, glare, and high contrast. Although other spectral bands like near-infrared (NIR) and long-wave infrared (LWIR) can enhance vision perception in such situations, they have limitations and lack large-scale datasets and benchmarks. Short-wave infrared (SWIR) imaging offers several advantages over NIR and LWIR. However, no publicly available large-scale datasets currently incorporate SWIR data for autonomous driving. To address this gap, we introduce the RGB and SWIR Multispectral Driving (RASMD) dataset, which comprises 100,000 synchronized and spatially aligned RGB-SWIR image pairs collected across diverse locations, lighting, and weather conditions. In addition, we provide a subset for RGB-SWIR translation and object detection annotations for a subset of challenging traffic scenarios to demonstrate the utility of SWIR imaging through experiments on both object detection and RGB-to-SWIR image translation. Our experiments show that combining RGB and SWIR data in an ensemble framework significantly improves detection accuracy compared to RGB-only approaches, particularly in conditions where visible-spectrum sensors struggle. We anticipate that the RASMD dataset will advance research in multispectral imaging for autonomous driving and robust perception systems.
The satire dataset is a new multi-modal dataset of satirical and regular news articles. The satirical news is collected from four websites that explicitly declare themselves to be satire, and the regular news is collected from six mainstream news websites. Specifically, the satirical news websites the articles were collected from are The Babylon Bee, Clickhole, Waterford Whisper News, and The DailyER. The regular news websites are Reuters, The Hill, Politico, New York Post, Huffington Post, and Vice News. The headlines and the thumbnail images of the latest 1000 articles for each of the publications are collected. The dataset contains a total of 4000 satirical and 6000 regular news articles. Source: A Multi-Modal Method for Satire Detection using Textual and Visual Cues
This is a dataset of 306,006 galaxies whose coordinates are taken from the Sloan Digital Sky Survey Data Release 7 and a modified catalogue from Brinchmann+2003 and Wilman+2010. This volume complete sample has an r-band absolute magnitude limit of $M_r\leq-20$ and a redshift limit of $z\leq0.08$. See Arora+2019 for details. This catalogue covers a wide range of environments from clusters to groups and field systems. The galaxy images are taken from the Dark Energy Spectroscopic Instrument, and contain $g$, $r$, and $z$ bands.
The SPOTS-10 dataset is an extensive collection of grayscale images showcasing diverse patterns found in ten animal species. Specifically, SPOTS-10 contains 50,000 32 x 32 grayscale images, divided into ten categories, with 5,000 images per category. The training set comprises 40,000 images, while the test set contains 10,000 images. The SPOTS-10 dataset is freely available on the project GitHub page: https://github.com/amotica/spots-10 by cloning the repository.
This dataset contains over 47,000 LEGO structures of over 28,000 unique 3D objects accompanied by detailed captions. It was used to train LegoGPT, the first approach for generating physically stable LEGO brick models from text prompts.
We introduce TextAtlas5M, a dataset specifically designed for training and evaluating multimodal generation models on dense-text image generation.
This dataset consists of 2,192 high-quality traditional Chinese landscape paintings (中国山水画). All paintings are sized 512x512, from the following sources: * Princeton University Art Museum, 362 paintings * Harvard University Art Museum, 101 paintings * Metropolitan Museum of Art, 428 paintings * Smithsonian's Freer Gallery of Art, 1,301 paintings Source: https://github.com/alicex2020/Chinese-Landscape-Painting-Dataset Image Source: https://github.com/alicex2020/Chinese-Landscape-Painting-Dataset
UDA-CH contains 16 objects that cover a variety of artworks which can be found in a museum like sculptures, paintings and books. Specifically, the dataset has been collected inside the cultural site “Galleria Regionale di Palazzo Bellomo” located in Siracusa, Italy. Source: An Unsupervised Domain Adaptation Scheme for Single-Stage Artwork Recognition in Cultural Sites
Inpainting networks are typically benchmarked on samples from Places2 dataset. However, this dataset does not have high resolution images for evaluation purposes. Instead, we will use images from the Unsplash-Lite Dataset, which contains 25k high resolution nature-themed photos. We randomly sampled 1000 images from the dataset. Each image is resized and cropped to 1024x1024, and a set of masks is generated with thin, medium, and thick brush strokes, using the methodology described in LaMa. The purpose of this dataset is to serve as a test-set for evaluating inpainting performance on high resolution natural images.
A high-resolution version of VGGFace2 for academic face editing purposes. This project uses GFPGAN for image restoration and insightface for data preprocessing (crop and align).
WiFiCam dataset for through-wall imaging based on WiFi channel state information. The corresponding source code repository is located at: https://github.com/StrohmayerJ/wificam Dataset Structure: /wificam ├── j3 └── 320 <-- 320x240 resolution subset └── csi.csv <-- raw WiFi packet sequence recorded with the ESP32-S3 └── csiComplex.npy <-- complex CSI sequence (cache) └── 92108.png <-- 320x240 RGB image └── 92112.png └── ... └── 640 <-- 640x480 resolution subset └── csi.csv <-- raw WiFi packet sequence recorded with the ESP32-S3 └── csiComplex.npy <-- complex CSI sequence (cache) └── 154.png <-- 640x480 RGB image └── 155.png └── ... ├── statistics320.csv <-- per-channel means and standard deviations for 320x240 images ├── statistics640.csv <-- per-channel means and standard deviations for 640x480 images
75k photos of windows + 21k synthetic renders of building windows.
A Filipino multi-modal language dataset for text+visual tasks. Consists of 351,755 Filipino news articles gathered from Filipino news outlets. Each entry contains: * body - Article text * title - Article title * website - Name of the news outlet * category - News category given by the news outlet * date - Date published * author - Article author * url - URL of the article * img_url - URL of the article image * img_path - Filename of the image in the dataset
This dataset is the images of corn seeds considering the top and bottom view independently (two images for one corn seed: top and bottom). There are four classes of the corn seed (Broken-B, Discolored-D, Silkcut-S, and Pure-P) 17802 images are labeled by the experts at the AdTech Corp. and 26K images were unlabeled out of which 9k images were labeled using the Active Learning (BatchBALD) We have created three different dataset: (1). Primary dataset: contains the 17802 images labeled by the experts. Top-view(8901) and Bottom-view(8901). (2). Dataset with fake images: We generated fake images using Conditional GAN (BigGAN) as follows: broken-2937, discolored-5823, pure-2937, silkcut-5823 instances and added them into the train set to balance the data set. (3). Balanced dataset: In this case of adding newly captured images labeled using the Batch Active Learning method, new 9000 labeled images are added into the primary dataset. This new dataset contains 26,802 images split into train and validation set 80: 20, respectively. Contains the 17802 images and the 9K images labeled by the Active Learning (BatchBALD).
The deliberate manipulation of public opinion, especially through altered images, poses a significant danger to society. To fight this issue on a technical level we support the research community by releasing the Digital Forensics 2023 (DF2023) training and validation dataset. The DF2023 training dataset comprises one million images from four major forgery categories: * splicing (400K) * copy-move (300K) * enhancement (200K) * removal (100K) This dataset enables an objective comparison of network architectures and can significantly reduce the time and effort of researchers preparing datasets. For a detailed description of the DF2023 dataset, please refer to: @inproceedings{Fischinger2023DFNet, title={DF2023: The Digital Forensics 2023 Dataset for Image Forgery Detection}, author={David Fischinger and Martin Boyer}, journal={The 25th Irish Machine Vision and Image Processing conference. (IMVIP)}, year={2023} } available from: Zenodo ## Naming convention The naming convention of DF2023 encodes information about the applied manipulations. Each image name has the following form: COCO_DF_0123456789_NNNNNNNN.{EXT} (e.g. COCO_DF_E000G40117_00200620.jpg) After the identifier of the image data source ("COCO") and the self-reference to the Digital Forensics ("DF") dataset, there are 10 digits as placeholders for the manipulation. Position 0 defines the manipulation types copy-move, splicing, removal, enhancement ([C,S,R,E]). The following digits 1-9 represent donor patch manipulations. For positions [1,2,7,8] (resample, flip, noise and brightness), a binary value indicates if this manipulation was applied to the donor image patch. Position 3 (rotate) indicates by the values 0-3 if the rotation was executed by 0, 90, 180 or 270 degrees. Position 4 defines if BoxBlur (B) or GaussianBlur (G) was used. Position 5 specifies the blurring radius. A value of 0 indicates that no blurring was executed. Position 6 indicates which of the Python-PIL contrast filters EDGE ENHANCE, EDGE ENHANCE MORE, SHARPEN, UnsharpMask or ImageEnhance (values 1-5) was applied. If none of them was applied, this value is set to 0. Finally, position 9 is set to the JPEG compression factor modulo 10, a value of 0 indicates that no JPEG compression was applied. The 8 characters NNNNNNNN in the image name template stand for a running number of the images.
Dataset for the DREAMING - Diminished Reality for Emerging Applications in Medicine through Inpainting Challenge! The DREAMING Challenge is part of ISBI 2024. Participate for Awards, Money Prices, Proceedings Paper, Top-Tier Challenge Summary Paper! More information about the challenge and registration can be found on the challenge platform! The DREAMING training dataset contains 100 diverse, photorealistic surgery scenes with unique patients, surroundings and obstructions from dynamically moving and static surgical instruments and hands holding them. We also provide the masks defining the area to be inpainted for each image.
The free Face dataset made for students and teachers. It contains 10,000 photos with equal distribution of race and gender parameters, along with metadata and facial landmarks. Free to use for research with citation Photos by Generated.Photos. Photos All the photos are 100% synthetic. Based on model-released photos. Royalty-free. Can be used for any research purpose except for the ones violating the law. Worldwide. No time limitations. Quantity 10,000 Quality 256x256px Diversity Ethnicity, gender Metadata The JSON files contain the metadata for each image in a machine-readable format, including: (1) FaceLandmarks: mouth, right_eyebrow, left_eyebrow, right_eye, left_eye, nose, jaw. (2) FaceAttributes: headPose, gender, makeup, emotion, facialHair, hair (hairColor, hairLength, bald), occlusion, ethnicity, eye_color, smile, age
InpaintCOCO is a benchmark to understand fine-grained concepts in multimodal models (vision-language) similar to Winoground. To our knowledge InpaintCOCO is the first benchmark, which consists of image pairs with minimum differences, so that the visual representation can be analyzed in a more standardized setting. A data sample contains 2 images and 2 corresponding captions that differ only in one object, the color of an object, or the size of an object. The metric used in the paper compares if true image-text pairs are more similar than wrong image-text combinations and that for both image-text pairs: \begin{equation} \begin{split} \operatorname{sim}(i_\text{COCO},t_\text{COCO}) > \operatorname{sim}(i_\text{inp},t_\text{COCO}) \quad \land \ \operatorname{sim}(i_\text{inp},t_\text{inp}) > \operatorname{sim}(i_\text{COCO},t_\text{inp}) \end{split} \label{eq:challengeset} \end{equation} InpaintCOCO is published in Proceedings of the 3rd Workshop on Advances in Language and Vision Research (ALVR) at ACL 2024.
Lemon dataset has been prepared to investigate the possibilities to tackle the issue of fruit quality control. It contains 2690 annotated images (1056 x 1056 pixels). Raw lemon images have been captured using the procedure described in the following blogpost and manually annotated using CVAT.
Multi-domain Image Editing Benchmark
NHR-Edit is a training dataset for instruction-based image editing. Each sample consists of an input image, a natural language editing instruction, and the corresponding edited image. All samples are generated fully automatically using the NoHumanRequired pipeline, without any human annotation or filtering. This dataset is designed to support training of general-purpose image editing models that can follow diverse, natural editing commands. Each sample also includes additional metadata such as editing type, style, and image resolution, making it suitable for training fine-grained, controllable image editing models.
We construct a fine-grained video-text dataset with 12K annotated high-resolution videos (~400k clips). The annotation of this dataset is inspired by the video script. If we want to make a video, we have to first write a script to organize how to shoot the scenes in the videos. To shoot a scene, we need to decide the content, shot type (medium shot, close-up, etc), and how the camera moves (panning, tilting, etc). Therefore, we extend video captioning to video scripting by annotating the videos in the format of video scripts. Different from the previous video-text datasets, we densely annotate the entire videos without discarding any scenes and each scene has a caption with ~145 words. Besides the vision modality, we transcribe the voice-over into text and put it along with the video title to give more background information for annotating the videos. In addition, we propose Vript-Bench, a new benchmark consisting of three challenging video understanding tasks (which are carefully double-checked by humans): - Vript-CAP (Caption): A benchmark with detailed captions rather than short captions. - Vript-RR (Retrieve then Reason): A video reasoning benchmark by first giving a detailed description of the scene as a hint and then asking questions about details in the scene. - Vript-ERO (Event Re-ordering): A benchmark that tests the temporal understanding by offering the descriptions of scenes located in two/four different timelines of the same video, and asks the model to give the right temporal order of the scenes.
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.