Context

Sample images for imaging experiments.

Copyright

Copyright is complicated, see here:

http://sipi.usc.edu/database/copyright.php

Content

Images kept in USC-SIPI Image Database

Acknowledgements

http://sipi.usc.edu/database/database.php

Inspiration

Image Processing algorithms created the need for sample images. They are kind of "vintage" now, but can still be used for interesting projects.

Images adapted from the MIT image registration course by Zöllei and Wells:https://ocw.mit.edu/courses/health-sciences-and-technology/hst-582j-biomedical-signal-and-image-processing-spring-2007/lecture-notes/l16_reg1.pdf

Sample royalty free images from https://www.pexels.com for various applications (computer vision, image processing, etc.)

This data contains X-ray computed tomography (XCT) reconstructed slices of additively manufactured cobalt chrome samples produced with varying laser powder bed fusion (LPBF) processing parameters (scan speed and hatch spacing). A constant laser power of 195 W and a layer thickness of 20 µm were used. Unoptimized processing parameters created defects in these parts. The as-built CoCr disks were 40 mm in diameter and 10 mm in height, with no post-processing step (e.g. heat treatment or hot isostatic pressing) used. Five mm diameter cylinders were cored out of each disk, and regions of interests (ROIs) within the cylinders were measured with XCT. The voxel size is approximately 2.5 µm, and approximately 1000 x 1000 x 1000 voxel three-dimensional images were obtained, for an actual volume of about (pi/4) x (2.5 mm)^3 in case of the approximately 2.5 µm voxel data sets. The data set contains two folders ('raw' and 'segmented') with 5 zipped tiff image folders, one for each sample. The images in the 'raw' folder are the original 16-bit XCT reconstructed images. The images in the 'segmented' folder are the segmented images. 'setn' in the file name represents the sample set and 'samplen' represents the sample number. The final trailing -n represents the number of the image in the stack where higher number is toward the top of the sample.

This dataset contains a collection of medical imaging files for use in the "Medical Image Processing with Python" lesson, developed by the Netherlands eScience Center.

The dataset includes:

1. SimpleITK compatible files: MRI T1 and CT scans (training_001_mr_T1.mha, training_001_ct.mha), digital X-ray (digital_xray.dcm in DICOM format), neuroimaging data (A1_grayT1.nrrd, A1_grayT2.nrrd). Data have been downloaded from here.
2. MRI data: a T2-weighted image (OBJECT_phantom_T2W_TSE_Cor_14_1.nii in NIfTI-1 format). Data have been downloaded from here.
3. Example images for the machine learning lesson: chest X-rays (rotatechest.png, other_op.png), cardiomegaly example (cardiomegaly_cc0.png).
4. Additional anonymized data: TBA

These files represent various medical imaging modalities and formats commonly used in clinical research and practice. They are intended for educational purposes, allowing students to practice image processing techniques, machine learning applications, and statistical analysis of medical images using Python libraries such as scikit-image, pydicom, and SimpleITK.

Introducing a comprehensive dataset designed to address the common challenge of pairing RGB images with their corresponding black and white (BW) versions. This meticulously curated dataset aims to provide a solution not only for personal use but also for researchers and developers encountering similar hurdles in their projects.

This dataset serves as a versatile resource applicable across a multitude of domains, including but not limited to image colorization models utilizing advanced techniques like Generative Adversarial Networks (GANs).

Organized for ease of use, the dataset comprises two distinct directories: one housing the RGB images and the other containing their corresponding BW counterparts. Each pair is conveniently named identically, facilitating effortless access to both versions of the same image.

Key Features:

- Comprehensive Pairing: Each RGB image is meticulously paired with its corresponding BW version, ensuring accuracy and reliability throughout the dataset.

- Clear Organization: With separate directories for RGB and BW images, users can effortlessly navigate and access the desired data without confusion.

- Ease of Accessibility: The consistent naming convention simplifies the process of retrieving both RGB and BW images, streamlining workflow for researchers and developers.

- Versatile Applications: The dataset's utility extends across various domains, from academic research to practical applications, fostering innovation and exploration in image processing and related fields.

How to Use:

Detailed documentation accompanies the dataset, providing comprehensive instructions on downloading, accessing, and leveraging the images effectively. Licensing:

The dataset is made available under Apache v2, ensuring clarity regarding its permissible usage, distribution, and modification. Sample Images:

A selection of sample image pairs showcases the dataset's contents and quality, offering users a glimpse into the richness of the data available. Dataset Statistics:

The dataset comprises 51289 pairs of RGB and BW images, totaling 102 578 in file size.

Community Engagement:

We encourage users to actively engage with the dataset, providing feedback, reporting issues, and sharing insights to foster collaboration and continuous improvement. This dataset represents a valuable contribution to the research and development community, offering a robust foundation for exploring and innovating in the realm of image processing and beyond. Harness its potential to unlock new possibilities and drive advancements in your projects.

Citation Request : 1. OZKAN IA., KOKLU M. and SARACOGLU R. (2021). Classification of Pistachio Species Using Improved K-NN Classifier. Progress in Nutrition, Vol. 23, N. 2, pp. DOI:10.23751/pn.v23i2.9686. (Open Access) 2. SINGH D, TASPINAR YS, KURSUN R, CINAR I, KOKLU M, OZKAN IA, LEE H-N., (2022). Classification and Analysis of Pistachio Species with Pre-Trained Deep Learning Models, Electronics, 11 (7), 981. https://doi.org/10.3390/electronics11070981. (Open Access) DATASET: https://www.muratkoklu.com/datasets/

https://www.mattioli1885journals.com/index.php/progressinnutrition/article/view/9686/9178 ABSTRACT: A computer vision system has been developed to distinguish two different species of pistachios with different characteristics that address different market types. 2148 sample image for these two kinds of pistachios were taken with a high-resolution camera. The image processing techniques, segmentation and feature extraction were applied on the obtained images of the pistachio samples. A pistachio dataset that has sixteen attributes was created. An advanced classifier based on k-NN method, which is a simple and successful classifier, and principal component analysis was designed on the obtained dataset. In this study; a multi-level system including feature extraction, dimension reduction and dimension weighting stages has been proposed. Experimental results showed that the proposed approach achieved a classification success of 94.18%. The presented high-performance classification model provides an important need for the separation of pistachio species and increases the economic value of species. In addition, the developed model is important in terms of its application to similar studies. Keywords: Classification, Image processing, k nearest neighbor classifier, Pistachio species

https://doi.org/10.3390/electronics11070981 Within the scope of the study, images of Kirmizi and Siirt pistachio types were obtained through the computer vision system. The pre-trained dataset includes a total of 2148 images, 1232 of Kirmizi type and 916 of Siirt type. Three different convolutional neural network models were used to classify these images. Models were trained by using the transfer learning method, with AlexNet and the pre-trained models VGG16 and VGG19. The dataset is divided as 80% training and 20% test. As a result of the performed classifications, the success rates obtained from the AlexNet, VGG16, and VGG19 models are 94.42%, 98.84%, and 98.14%, respectively. Models’ performances were evaluated through sensitivity, specificity, precision, and F-1 score metrics. In addition, ROC curves and AUC values were used in the performance evaluation. The highest classification success was achieved with the VGG16 model. The obtained results reveal that these methods can be used successfully in the determination of pistachio types. View Full-Text Keywords: pistachio; genetic varieties; machine learning; deep learning; food recognition

Overview of the morphometric investigations for sample images in Fig 7.

Image of single spheroids, to be used with the first generation macro (https://figshare.com/s/32f81784ee28e3fde015) for separate images arranged in a folder.

Here is a short description of a project that uses OpenCV in Python to process images:

Load an image: You can use the cv2.imread function to load an image into memory.

Pre-processing: Depending on your requirements, you can perform various operations on the image such as resizing, cropping, thresholding, or color conversion.

Image analysis: You can apply various image processing techniques like edge detection, feature extraction, object detection, or template matching to the pre-processed image.

Display results: After processing the image, you can display the results using the cv2.imshow function. You can also save the results to an image file using the cv2.imwrite function.

This project can be used for various applications such as object recognition, facial detection, image segmentation, and more. The implementation details and the algorithms used depend on the specific requirements of the project.

The LOCBEEF dataset contains 3268 images of local Aceh beef collected from 07:00 a.m - 22:00 p.m, more information about the clock is shown in Fig. The dataset contains two categories of directories, namely train, and test. Furthermore, each subdirectory consists of fresh and rotten. An example of the image can be seen in Figs. 2 and 3. The directory structure for the data is shown in Fig. 1. The image directory for train contains 2228 images each subdirectory contains 1114 images, and the test directory contains 980 images for each subdirectory containing 490 images. For images have a resolution of 176 x 144 pixel, 320 x 240 pixel, 640 x 480 pixel, 720 x 480 pixel, 720 x 720 pixel, 1280 x 720 pixel, 1920 x 1080 pixel, 2560 x 1920 pixel, 3120 x 3120 pixel, 3264 x 2248 pixel, and 4160 x 3120 pixel.

The classification of LOCBEEF datasets has been carried out using the deep learning method of Convolutional Neural Networks with an image composition of 70% training data and 30% test data. Images with the mentioned dimensions are included in the LOCBEEF dataset to apply to the Resnet50.

RESEARCH APPROACH

The research approach adopted for the study consists of seven phases which includes as shown in Figure 1:

Pre-acquisition

data pre-processing

Raw images collection

Image pre-processing

Naming of images

Dataset Repository

Performance Evaluation

The different phases in the study are discussed in the sections below.

PRE-ACQUISITION

The volunteers are given brief orientation on how their data will be managed and used for research purposes only. After the volunteers agrees, a consent form is given to be read and signed. The sample of the consent form filled by the volunteers is shown in Figure 1.

The capturing of images was started with the setup of the imaging device. The camera is set up on a tripod stand in stationary position at the height 90 from the floor and distance 20cm from the subject.

EAR IMAGE ACQUISITION

Image acquisition is an action of retrieving image from an external source for further processing. The image acquisition is purely a hardware dependent process by capturing unprocessed images of the volunteers using a professional camera. This was acquired through a subject posing in front of the camera. It is also a process through which digital representation of a scene can be obtained. This representation is known as an image and its elements are called pixels (picture elements). The imaging sensor/camera used in this study is a Canon E0S 60D professional camera which is placed at a distance of 3 feet form the subject and 20m from the ground.

This is the first step in this project to achieve the project’s aim of developing an occlusion and pose sensitive image dataset for black ear recognition. (OPIB ear dataset). To achieve the objectives of this study, a set of black ear images were collected mostly from undergraduate students at a public University in Nigeria.

The image dataset required is captured in two scenarios:

1. uncontrolled environment with a surveillance camera

The image dataset captured is purely black ear with partial occlusion in a constrained and unconstrained environment.

1. controlled environment with professional cameras

The ear images captured were from black subjects in controlled environment. To make the OPIB dataset pose invariant, the volunteers stand on a marked positions on the floor indicating the angles at which the imaging sensor was captured the volunteers’ ear. The capturing of the images in this category requires that the subject stand and rotates in the following angles 60o, 30o and 0o towards their right side to capture the left ear and then towards the left to capture the right ear (Fernando et al., 2017) as shown in Figure 4. Six (6) images were captured per subject at angles 60o, 30o and 0o for the left and right ears of 152 volunteers making a total of 907 images (five volunteers had 5 images instead of 6, hence folders 34, 22, 51, 99 and 102 contain 5 images).

To make the OPIB dataset occlusion and pose sensitive, partial occlusion of the subject’s ears were simulated using rings, hearing aid, scarf, earphone/ear pods, etc. before the images are captured.

CONSENT FORM

This form was designed to obtain participant’s consent on the project titled: An Occlusion and Pose Sensitive Image Dataset for Black Ear Recognition (OPIB). The information is purely needed for academic research purposes and the ear images collected will curated anonymously and the identity of the volunteers will not be shared with anyone. The images will be uploaded on online repository to aid research in ear biometrics.

The participation is voluntary, and the participant can withdraw from the project any time before the final dataset is curated and warehoused.

Kindly sign the form to signify your consent.

I consent to my image being recorded in form of still images or video surveillance as part of the OPIB ear images project.

Tick as appropriate:

GENDER Male Female

AGE (18-25) (26-35) (36-50)

………………………………..

SIGNED

Figure 1: Sample of Subject’s Consent Form for the OPIB ear dataset

RAW IMAGE COLLECTION

The ear images were captured using a digital camera which was set to JPEG because if the camera format is set to raw, no processing will be applied, hence the stored file will contain more tonal and colour data. However, if set to JPEG, the image data will be processed, compressed and stored in the appropriate folders.

IMAGE PRE-PROCESSING

The aim of pre-processing is to improve the quality of the images with regards to contrast, brightness and other metrics. It also includes operations such as: cropping, resizing, rescaling, etc. which are important aspect of image analysis aimed at dimensionality reduction. The images are downloaded on a laptop for processing using MATLAB.

Image Cropping

The first step in image pre-processing is image cropping. Some irrelevant parts of the image can be removed, and the image Region of Interest (ROI) is focused. This tool provides a user with the size information of the cropped image. MATLAB function for image cropping realizes this operation interactively by waiting for a user to specify the crop rectangle with the mouse and operate on the current axes. The output images of the cropping process are of the same class as the input image.

Naming of OPIB Ear Images

The OPIB ear images were labelled based on the naming convention formulated from this study as shown in Figure 5. The images are given unique names that specifies the subject, the side of the ear (left or right) and the angle of capture. The first and second letters (SU) in the image names is block letter simply representing subject for subject 1-to-n in the dataset, while the left and right ears is distinguished using L1, L2, L3 and R1, R2, R3 for angles 600, 300 and 00, respectively as shown in Table 1.

Table 1: Naming Convention for OPIB ear images

NAMING CONVENTION

Label

Degrees 600 300 00

No of the degree 1 2 3

Subject 1 indicates (first image in dataset) SU1

Subject n indicates (last image in dataset) SUn

Left Image 1 L 1

Left image n L n

Right Image 1 R 1

Right Image n R n

SU1L1 SU1RI

SU1L2 SU1R2

SU1L3 SU1R3

OPIB EAR DATASET EVALUATION

The prominent challenges with the current evaluation practices in the field of ear biometrics are the use of different databases, different evaluation matrices, different classifiers that mask the feature extraction performance and the time spent developing framework (Abaza et al., 2013; Emeršič et al., 2017).

The toolbox provides environment in which the evaluation of methods for person recognition based on ear biometric data is simplified. It executes all the dataset reads and classification based on ear descriptors.

DESCRIPTION OF OPIB EAR DATASET

OPIB ear dataset was organised into a structure with each folder containing 6 images of the same person. The images were captured with both left and right ear at angle 0, 30 and 60 degrees. The images were occluded with earing, scarves and headphone etc. The collection of the dataset was done both indoor and outdoor. The dataset was gathered through the student at a public university in Nigeria. The percentage of female (40.35%) while Male (59.65%). The ear dataset was captured through a profession camera Nikon D 350. It was set-up with a camera stand where an individual captured in a process order. A total number of 907 images was gathered.

The challenges encountered in term of gathering students for capturing, processing of the images and annotations. The volunteers were given a brief orientation on what their ear could be used for before, it was captured, for processing. It was a great task in arranging the ear (dataset) into folders and naming accordingly.

Table 2: Overview of the OPIB Ear Dataset

Location

Both Indoor and outdoor environment

Information about Volunteers

Students

Gender

Female (40.35%) and male (59.65%)

Head Side Left and Right

Side Left and Right

Total number of volunteers

152

Per Subject images

3 images of left ear and 3 images of right ear

Total Images

907

Age group

18 to 35 years

Colour Representation

RGB

Image Resolution

224x224

Execution time comparison of different schemes for sample images.

This extensive dataset presents a meticulously curated collection of low-resolution images showcasing 20 well-established rice varieties native to diverse regions of Bangladesh. The rice samples were carefully gathered from both rural areas and local marketplaces, ensuring a comprehensive and varied representation. Serving as a visual compendium, the dataset provides a thorough exploration of the distinct characteristics of these rice varieties, facilitating precise classification.

Dataset Composition

The dataset encompasses 20 distinct classes, encompassing Subol Lota, Bashmoti (Deshi), Ganjiya, Shampakatari, Sugandhi Katarivog, BR-28, BR-29, Paijam, Bashful, Lal Aush, BR-Jirashail, Gutisharna, Birui, Najirshail, Pahari Birui, Polao (Katari), Polao (Chinigura), Amon, Shorna-5, and Lal Binni. In total, the dataset comprises 4,730 original JPG images and 23,650 augmented images.

Image Capture and Dataset Organization

These images were captured using an iPhone 11 camera with a 5x zoom feature. Each image capturing these rice varieties was diligently taken between October 18 and November 29, 2023. To facilitate efficient data management and organization, the dataset is structured into two variants: Original images and Augmented images. Each variant is systematically categorized into 20 distinct sub-directories, each corresponding to a specific rice variety.

Original Image Dataset

The primary image set comprises 4,730 JPG images, uniformly sized at 853 × 853 pixels. Due to the initial low resolution, the file size was notably 268 MB. Employing compression through a zip program significantly optimized the dataset, resulting in a final size of 254 MB.

Augmented Image Dataset

To address the substantial image volume requirements of deep learning models for machine vision, data augmentation techniques were implemented. Total 23,650 images was obtained from augmentation. These augmented images, also in JPG format and uniformly sized at 512 × 512 pixels, initially amounted to 781 MB. However, post-compression, the dataset was further streamlined to 699 MB.

Dataset Storage and Access

The raw and augmented datasets are stored in two distinct zip files, namely 'Original.zip' and 'Augmented.zip'. Both zip files contain 20 sub-folders representing a unique rice variety, namely 1_Subol_Lota, 2_Bashmoti, 3_Ganjiya, 4_Shampakatari, 5_Katarivog, 6_BR28, 7_BR29, 8_Paijam, 9_Bashful, 10_Lal_Aush, 11_Jirashail, 12_Gutisharna, 13_Red_Cargo,14_Najirshail, 15_Katari_Polao, 16_Lal_Biroi, 17_Chinigura_Polao, 18_Amon, 19_Shorna5, 20_Lal_Binni.

Train and Test Data Organization

To ease the experimenting process for the researchers we have balanced the data and split it in an 80:20 train-test ratio. The ‘Train_n_Test.zip’ folder contains two sub-directories: ‘1_TEST’ which contains 1125 images per class and ‘2_VALID’ which contains 225 images per class.

• Collecting Images Images of the surface roughness were captured with the use of a Nikon Eclipse L150 Optical Microscope and a Clemex Image Analysis System. Using a camera, we were able to quickly upload the images that we captured with the microscope onto the computer. The model of the camera that was utilized is known as the Sentech USB 2.0 (Triggered) Color Camera STC-TC202USB-AH. The images captured by the camera were processed with software called Clemex Vision PE (V7). Only for image collection and samples with a high surface roughness can this software make advantage of its multi-layer Grab characteristics, which allow for alternative image processing processes.

• Dataset Details Sample images were taken at 50X, 100X, and 200X magnification for image processing purposes. The captured images' resolution ranges from 1600 to 1200 pixels, and the file sizes range from roughly 3-4 Mb in size. A minimum of 6 different images were obtained from each class at various levels of magnification. For instance, the construction of the 100X dataset required the use of a total of 288 (16x6x3) high-resolution images. With the use of cropping, we were able to obtain 40 images from each high-resolution image. Hence, a class consists of 240 images. In total, we looked at 3840 images (240*16) for 16 different classes.

• Classification Details Three different surface roughness sets are utilized as 50X, 100X, and 200X categories. The 50X dataset contains 4320 samples, whereas both the 200X and 100X dataset includes 3840 samples. Both datasets have 16 classes.

• Cite Anagun, Y, Isik, S and Cakir, F.H. "Surface Roughness Classification of Electro Discharge Machined Surfaces with Deep Ensemble Learning." Measurement (2023): 112855, doi: 10.1016/j.measurement.2023.112855.

Link: https://www.sciencedirect.com/science/article/pii/S0263224123004190

• License Free only for research purposes, for commercial use, please contact us(drsahinisik@gmail.com).

This dataset is composed of 81 pairs of correlated images. Each pair contains one image of an iron ore sample acquired through reflected light microscopy (RGB, 24-bit), and the corresponding binary reference image (8-bit), in which the pixels are labeled as belonging to one of two classes: ore (0) or embedding resin (255).

The sample came from an itabiritic iron ore concentrate from Quadrilátero Ferrífero (Brazil) mainly composed of hematite and quartz, with little magnetite and goethite. It was classified by size and concentrated with a dense liquid. Then, the fraction -149+105 μm with density greater than 3.2 was cold mounted with epoxy resin and subsequently ground and polished.

Correlative microscopy was employed for image acquisition. Thus, 81 fields were imaged on a reflected light microscope with a 10× (NA 0.20) objective lens and on a scanning electron microscope (SEM). In sequence, they were registered, resulting in images of 999×756 pixels with a resolution of 1.05 µm/pixel. Finally, the images from SEM were thresholded to generate the reference images.

Further description of this sample and its imaging procedure can be found in the work by Gomes and Paciornik (2012).

This dataset was created for developing and testing deep learning models on semantic segmentation tasks. The paper of Filippo et al. (2021) presented a variant of the DeepLabv3+ model that reached mean values of 91.43% and 93.13% for overall accuracy and F1 score, respectively, for 5 rounds of experiments (training and testing), each with a different, random initialization of network weights.

For further questions and suggestions, please do not hesitate to contact us.

Contact email: ogomes@gmail.com

If you use this dataset in your own work, please cite this DOI: 10.5281/zenodo.5014700

Please also cite this paper, which provides additional details about the dataset:

Michel Pedro Filippo, Otávio da Fonseca Martins Gomes, Gilson Alexandre Ostwald Pedro da Costa, Guilherme Lucio Abelha Mota. Deep learning semantic segmentation of opaque and non-opaque minerals from epoxy resin in reflected light microscopy images. Minerals Engineering, Volume 170, 2021, 107007, https://doi.org/10.1016/j.mineng.2021.107007.

This dataset includes photographic images of thin sections created from hand samples and drill core collected from zoned ultramafic intrusions in Alaska, a shapefile representing the locations of the hand samples and drill core collars, and images showing examples of the approximate scale on the images. The samples were collected in order to help define nickel, copper, platinum group element (PGE), gold, and titanium vanadium iron resources associated with the intrusions. The images of the entire thin section, in plane-polarized (PPL) and cross-polarized light (XPL), were taken using a high-resolution digital camera on a macro stand. The PPL and XPL images were exported from the camera in both raw .CR2 and compressed .jpg format. Reflected light (RL) images of the thin sections were taken using a Keyence VHX-7000 digital microscope. The RL images were exported from the Keyence as .tif images. The data are organized into four zip files, three containing the thin section images ((AK_ZUM_thin_section_images_part1.zip; (AK_ZUM_thin_section_images_part2.zip; (AK_ZUM_thin_section_images_part3.zip), and one containing a shapefile of the locations of the hand samples (AK_ZUM_thinsection_sample_locations_shapefile.zip). The image files were divided into three zip files due to file size upload restrictions. The thin section images within each zip file are in two folders, one containing the 'RAW' unprocessed .CR2 images and associated .xmp metadata files, and unprocessed .tif images (AK_ZUM_Original_Images_part'x'), and another folder containing processed .jpg images (AK_ZUM_Processed_Images_part'x'). The processed images were run through a series of tools in Photoshop in order to improve the appearance of the images. The images taken in reflected light were not run through the Photoshop processing. Hand sample identifiers (ID's) include a two-digit number representing the year they were collected, an abbreviated location identifier, and a random sample number. Images with file names containing sample ID's with "_A" or "_B" indicate two thin sections sourced from the same sample. The hand sample location abbreviations and their meanings are: BI=Blashke Island, DI=Duke Island, HUM=Haines, KP=Kane Peak, KUM=Klukwan, SI=Sukoi Island, PSUM=Port Snettisham, UB=Union Bay. The drill hole location abbreviations and their meanings are: DK=Duke Island, KW=Klukwan, SC=Salt Chuck, SN=Port Snettisham, UB=Union Bay. The file name of each photograph correlates to either the drill hole and depth of the sample, or the hand sample ID from which the thin section was sourced, and whether the photograph represents a plane-polarized (PPL), cross-polarized (XPL), or reflected (RL) light image. For example, the "SN01_395-396_XPL.JPG" file is a cross-polarized image of the thin section sourced from a sample taken from the "SN01" drill hole at a depth of 395-396 feet. The RL images include a scale bar. The PPL and XPL images do not show a scale bar, however the thin section holder, viewable in the images, can be used as an approximate scale bar. The width of the rectangular opening of the thin section holder is approximately 48 millimeters wide. See the "Scale_bar_example_PPL.JPG" and "Scale_bar_example_XPL.JPG" image files for examples. The image files have Section 508 compliant metadata as per USGS section 508 compliancy officer guidelines. The .xmp files are included with the 'RAW' .CR2 files because the original .CR2 metadata cannot be altered. In order to add descriptive information to the metadata, the .xmp files must be included. The image metadata for the .jpg and .tif files can be viewed using software such as Adobe Photoshop, Adobe Bridge, or Windows file explorer. The raw .CR2 images edited metadata (.xmp files) can be viewed in Adobe Photoshop and Adobe Bridge, but is not viewable using Windows file explorer. For the .CR2 images, only the original metadata, created when the photograph was taken, can be be viewed using Windows file explorer.

Overview: The Image and Text Pair Dataset is a curated collection of images paired with descriptive textual captions or subtitles. This dataset is designed to support various natural language processing and computer vision tasks, such as image captioning, text-to-image retrieval, and multimodal machine learning research. It serves as a valuable resource for training and evaluating models that can understand and generate meaningful relationships between visual content and textual descriptions.

Contents: The dataset consists of the following components:

Images: The dataset includes a set of image files in common formats such as JPEG or PNG. Each image captures a different scene, object, or concept. These images are diverse and cover a wide range of visual content.

Textual Captions or Subtitles: For each image, there is an associated textual caption or subtitle that describes the content of the image. These captions provide context, details, or descriptions of the visual elements in the images. The text data is in natural language and is designed to be human-readable.

Use Cases: The Image and Text Pair Dataset can be utilized for various machine learning and AI tasks, including but not limited to:

Image Captioning: Training and evaluating models to generate textual descriptions for given images. Text-to-Image Retrieval: Enabling models to retrieve images based on textual queries. Multimodal Learning: Supporting research in multimodal models that understand and bridge the gap between textual and visual data. Natural Language Processing: Serving as a source of textual data for NLP tasks like text generation, summarization, and sentiment analysis. Dataset Size: The dataset contains a specific number of image and text pairs. The exact number may vary depending on the dataset's source and purpose. It may range from a few dozen pairs to thousands or more, depending on its intended application.

Data Sources: The source of this dataset may vary. In this case, the images and captions have been uploaded to a platform like Kaggle. They could be sourced from a variety of places, including user-generated content, public image collections, or custom data creation.

Research and Applications: Researchers and practitioners can use this dataset to advance the state of the art in various AI fields, particularly in areas where understanding and generating text-image relationships are critical. It can be a valuable resource for building models that can comprehend and describe visual content, as well as for developing innovative applications in areas like image recognition, image search, and content recommendation.

Please note that the specifics of the dataset, including the number of image-caption pairs, data sources, and licensing, can vary depending on the actual dataset you have uploaded to Kaggle or any other platform. The above description is a generalized template and can be adapted to your specific dataset's details.

Paired data of z-sweeps and z-projections of 3D fluorescent samples.Note: The "E**.tif" images are excluded due to saturated fluorescent signals

Introduction This is my scraped, collected, and curated dataset for the Google Universal Image Embedding competition resized to 512x512. It contains 130k+ images in total and below provides a count for each class -

Data Count | apparel | artwork | cars | dishes | furniture | illustrations | landmark | meme | packaged | storefronts | toys | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | 32,226| 4,957 | 8,144 | 5,831 | 10,488 | 3,347 | 33,063 | 3,301 | 23,382 | 5,387 | 2,402 |

Data Source 1. Apparel - Deep Fashion Dataset 2. Artwork - Google Scrapped 3. Cars - Stanford Cars Dataset 4. Dishes - Google Scrapped 5. Furniture - Google Scrapped 6. Illustrations - Google Scrapped 7. Landmark - Google Landmark Dataset 8. Meme - Google Scrapped 9. Packaged - Holosecta, Grozi 3.2k, Freiburg Groceries, SKU110K 10. Storefronts - Google Scrapped 11. Toys - Google Scrapped