About this dataset

Do you want to help a dog in need? This dataset contains information on over 3,000 adoptable dogs across the United States. By understanding patterns of dog movement and relocation, we can help these animals find their forever homes.

The data includes information on the origin of each dog, as well as the state they are currently listed for adoption in. This can be used to understand patterns of dog movement across the country, and how different states rely on imported dogs for adoption.

There are several things to keep in mind when using this dataset: - The data represents a single day of data. It is possible that patterns have changed since then. - The data only includes adoptable dogs that were listed on PetFinder.com

How to use the dataset

This dataset of adoptable dogs in the US was collected to better understand how animals are relocated from state to state and imported from outside the US. The data includes information on over 3,000 dogs that were described as having originated in places different from where they were listed for adoption. The findings were published in a visual essay on The Pudding entitled Finding Forever Homes published in October 2019.

This dataset is a snapshot of data collected on a single day and does not include all adoptable dogs in the US. However, it provides valuable insights into the whereabouts of these animals and the journey they take to find their forever homes

So, how should you use it?

This dataset is a great resource for understanding how adoptable dogs are relocated from state to state and imported into the US. The data provides information on the origin of each dog, as well as the state they are currently listed for adoption in. This can be used to understand patterns of dog movement across the country, and how different states rely on imported dogs for adoption.

Research Ideas

• Finding out how many of each type and breed of dog are brought into shelters across the USA in a given year.
• Seeing which states have the most imports of dogs and what breeds/types those are.
• Determining if there are any trends in the types/breeds of dogs being brought into shelters (e.g. more pit bulls than golden retrievers)

Columns

File: dogTravel.csv | Column name | Description | |:------------------|:---------------------------------------------------------------------| | contact_city | The city where the animal is located. (String) | | contact_city | The city where the animal is located. (String) | | contact_state | The state where the animal is located. (String) | | contact_state | The state where the animal is located. (String) | | description | A description of the animal. (String) | | description | A description of the animal. (String) | | found | The date the animal was found. (Date) | | found | The date the animal was found. (Date) | | manual | A manual override for the animal's location. (String) | | manual | A manual override for the animal's location. (String) | | remove | The date the animal was removed from the dataset. (Date) | | remove | The date the animal was removed from the dataset. (Date) | | still_there | Whether or not the animal is still available for adoption. (Boolean) | | still_there | Whether or not the animal is still available for adoption. (Boolean) |

File: allDogDescriptions.csv | Column name | Description | |:--------------------|:-------------------------------------------------------| | contact_city | The city where the animal is located. (String) | | contact_city | The city where the animal is located. (String) | | contact_state | The state where the animal is located. (String) | | contact_state | The state where the animal is located. (String) | | description | A description of the animal. (String) | | description | A description of the animal. (String) | | url | The URL of the animal's profile on PetFinder. (String) | | url | The URL of the animal's profile on PetFinder. (String) | | type.x | The type of animal. (String) | | type.x | The type of animal. (String) | | species | The species of the animal. (S...

Dog Person

## Overview

Dog Person is a dataset for object detection tasks - it contains Dogs Cats Person annotations for 2,574 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).

Context

This dataset contains images of 120 breeds of dogs from around the world. This dataset has been built using images and annotation from ImageNet for the task of fine-grained image categorization. It was originally collected for fine-grain image categorization, a challenging problem as certain dog breeds have near identical features or differ in colour and age.

Content

• Number of categories: 120
• Number of images: 20,580
• Annotations: Class labels, Bounding boxes

Acknowledgements

The original data source is found on http://vision.stanford.edu/aditya86/ImageNetDogs/ and contains additional information on the train/test splits and baseline results. If you use this dataset in a publication, please cite the dataset on the following papers: Aditya Khosla, Nityananda Jayadevaprakash, Bangpeng Yao and Li Fei-Fei. Novel dataset for Fine-Grained Image Categorization. First Workshop on Fine-Grained Visual Categorization (FGVC), IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2011. Secondary: J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li and L. Fei-Fei, ImageNet: A Large-Scale Hierarchical Image Database. IEEE Computer Vision and Pattern Recognition (CVPR), 2009.

Inspiration

• Can you correctly identify dog breeds that have similar features, such as the basset hound and bloodhound?
• Is this chihuahua young or old?

Dataset Card for Cats Vs. Dogs

  Dataset Summary

A large set of images of cats and dogs. There are 1738 corrupted images that are dropped. This dataset is part of a now-closed Kaggle competition and represents a subset of the so-called Asirra dataset. From the competition page:

The Asirra data set Web services are often protected with a challenge that's supposed to be easy for people to solve, but difficult for computers. Such a challenge is often called a CAPTCHA… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/cats_vs_dogs.

Dataset used in the article "Does Visual Stimulation by Photographs of Cats and Dogs Make People Happier and More Optimistic?"ColumnsIDis_preview: true - response by the researcher to check the questionnaire, it should be removedremove: respondent checked that his/her responses are not valid and should not be used in future analysisfinished_proc: percentage of the questionnaire finisheddate_time: filing of the questionnaire started at this timeduration_formatted: duration of the filling of the questionnairebrowserbrowser_versionOS: operating systempriming: true - primed group, false - control groupcat_dog: objects on photos showngenderage: in yerssex_o: attraction to people of the opposite sex (scale 1 - 7)sex_s: attraction to people of the same sex (scale 1 - 7) orientation: computed as the difference of previous twomood: actual mood (scale 0 - 5)condition_phys: physical condition (scale 0 - 5)condition_psych: mental condition (scale 0 - 5)life_quality: life quality (scale 0 - 5)optimism: mean of previous threeoptimism_zskore: z-score of the previous children_own: how many children does respondent havewanted_sons: total number of sons which respondent would like to havewanted_daughters: total number of daughters which respondent would like to havewanted_children: a sum of previous twoliking_dogs: how much respondent likes dogs (scale 1 - 100)present_whenever_dog: respondent has ever kept a dogpresent_now_dog: respondent keeps dog nowpresent_Ndogs: how many dogs does respondent keep now liking_cats: how much respondent likes cats (scale 1 - 100)present_whenever_cat: respondent has ever kept a catpresent_now_cat: respondent keeps cat nowpresent_Ncats: how many cats does respondent keep now

ProcedureWe have conducted two surveys in Germany, both were developed by Jesko Wilke, a freelancer journalist of the German ‘Dogs’ magazine. The data were collected online by the magazine’s own website (www.dogs-magazin.de). The surveys were described in detail in Kubinyi et al., 2009; Turcsán et al., 211 and Turcsán et al., 2017. Both surveys comprised two parts. The first part collected information about the demographic characteristics of the owners and dogs, as well as about the dog keeping practices. Twelve of these questions were the same in both surveys, eight were present in only one. The second part was different in the two surveys. The Survey 1 aimed at measuring the dogs’ general behaviour tendencies (personality) and was developed based on a human Big Five Inventory. This questionnaire contained 24 items (e.g. „My dog is calm, even in ambiguous situations”), for each item the owners were asked to indicate the level of agreement on a 3-point scale (true, partly true, not true). Our previous results using principal component analysis have revealed that 17 items out of the 24 belonged to four components, labelled as calmness, trainability, dog sociability, and boldness, all traits with middle or high internal consistency.The Survey 2 listed 12 examples of typical behaviour problems like „ My dog most often does not even attend me when I call him/her back”. Again, the owners indicated for each statement how far they agree with it using a 3-point scale. The questions were designed to assess not (only) the frequency of behaviour problems of the dogs but (also) the owners’ attitude towards these behaviour; i.e. if he/she considers them as problematic. In the current dataset, we recoded responses into a binary (yes/no) format: responses of "agree" or "partly agree" were categorized as "yes", while "disagree" was categorized as "no".SubjectsOn total, we collected responses from N = 14,004 dog owners in the first survey and N = 10,240 in the second. In the current dataset, we excluded reports with- missing data- duplicate entries (i.e., cases where owners submitted multiple reports for the same dog)- reports on mixed-breed dogs- reports on breeds where the cephalic index of the breed was unknown- reports when the cephalic index of the breed fell between 50 and 53, and between 62 and 65.Finally, to prevent a few highly popular breeds from disproportionately influencing group values, we capped the number of individuals per breed at 100. If a breed exceeded this threshold, we randomly selected 100 individuals for the final dataset.Kubinyi, E., Turcsán, B. & Miklósi, Á. Dog and owner demographic characteristics and dog personality trait associations. Behavioural Processes 81, 392–401 (2009).Turcsán, B., Kubinyi, E. & Miklósi, Á. Trainability and boldness traits differ between dog breed clusters based on conventional breed categories and genetic relatedness. Applied Animal Behaviour Science 132, 61–70 (2011).Turcsán, B., Miklósi, Á. & Kubinyi, E. Owner perceived differences between mixed-breed and purebred dogs. PLoS ONE 12, (2017).

This dataset is about books. It has 2 rows and is filtered where the book is The people with the dogs. It features 7 columns including author, publication date, language, and book publisher.

The "Cluster Analysis of Pet Owners" dataset, consisting of 250 entries, provides a detailed view of various dimensions of pet ownership. It contains Likert scale items answered from 1 - Strongly Disagree to 4 - Strongly Agree. It includes personal assessments of the impact pets have on owners' well-being, with statements like "Owning a pet has helped my health" and "Owning a pet adds to my happiness." Additionally, it captures attachment levels and the emotional bonds owners feel toward their pets through statements such as "I am very attached to my pet" and "My pet and I have a close relationship." This dimension reflects how pet ownership affects emotional well-being and connection, critical for understanding the strength of these owner-pet relationships.

Beyond emotional bonds, the dataset explores the interaction frequency and nature between owners and pets, such as through statements like "I play with my pet quite often" and "I often take my pet along when I visit friends." A separate set of variables examines companionship, with items like "My pet is like a friend that can keep me from being lonely," highlighting pets' social and emotional roles. Furthermore, the dataset includes Recency, Frequency, and Monetary (RFM) metrics, likely indicating recent engagement levels, frequency of interaction, and expenditures on pets. This mix of emotional, social, and financial metrics provides a rich basis for clustering pet owners based on their behaviors, attachment levels, and perceived benefits of pet ownership.

All dog owners residing in NYC are required by law to license their dogs. The data is sourced from the DOHMH Dog Licensing System, where owners can apply for and renew dog licenses.

Each record represents a unique dog license that was active during the year, but not necessarily a unique record per dog, since a license that is renewed during the year results in a separate record of an active license period. Each record stands as a unique license period for the dog over the course of the yearlong time frame.

This dataset is useful for municipal governments, veterinarians, and researchers who are interested in pet ownership patterns, compliance with local licensing laws, and demographic analysis of pet ownership. It can also aid in public health monitoring, such as tracking rabies vaccinations, which are often required for licensing.

Data scientists and analysts can perform various types of analytics such as:

• Descriptive Statistics: Analyze the distribution of dog breeds, the average age of dogs, and the licensing compliance rates across different zip codes.
• Geospatial Analysis: Map the data to see geographic patterns in pet ownership, breed popularity, or areas with lower licensing compliance, which can help in targeted public health and safety campaigns.
• Time Series Analysis: If longitudinal data is available, trends over time in pet registrations, renewals, and compliance could be evaluated.
• Predictive Modeling: Model the likelihood of license renewal or the factors influencing it, using logistic regression or other suitable predictive models.
• Cluster Analysis: Identify clusters of similar zip codes based on dog population characteristics, which can help in resource allocation for animal control agencies.

Dogs and cats have become the most important and successful pets through long-term domestication. People keep them for various reasons, such as their functional roles or for physical or psychological support. However, why humans are so attached to dogs and cats remains unclear. A comprehensive understanding of the current state of human preferences for dogs and cats and the potential influential factors behind it is required. Here, we investigate this question using two independent online datasets and anonymous questionnaires in China. We find that current human preferences for dog and cat videos are relatively higher than for most other interests, with video plays ranking among the top three out of fifteen interests. We also find genetic variations, gender, age, and economic development levels notably influence human preferences for dogs and cats. Specifically, dog and cat ownership are significantly associated with parents’ pet ownership of dogs and cats (Spearman’s rank correlation c..., , , # Human preferences for dogs and cats in China: the current situation and influencing factors of watching online videos and pet ownership

https://doi.org/10.5061/dryad.qfttdz0rr

This dataset contains three CSV data files, each corresponding to one of the three parts described in the study.

Description of the data and file structure

**“1, bilibili.csv†**: contains data extracted from the Bilibili website. Each row in the dataset represents yearly data for each popular channel. Missing data are indicated with NA.

• ID:Â The serial number for each video, ranging from 1 to 167368.
• year: The year the video was published on the website, from 2009 to 2021.
• Videourl:Â The URL of the video.
• plays:Â The total number of plays for the video.
• likes: The total number of likes for the video.
• sort: The ranking of the video in terms of play count among all popular videos in its channel for that year.
• channelID: The I...

Thermal People And Dogs

## Overview

Thermal People And Dogs is a dataset for object detection tasks - it contains Thermal annotations for 619 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 dataset is a modelled dataset, describing the predicted population of dogs per postcode district (e.g. YO41). This dataset gives the mean estimate for population for each district, and was generated as part of the delivery of commissioned research. The data contained within this dataset are modelled figures, based on national estimates for pet population, and available information on Veterinary activity across GB. The data are accurate as of 01/01/2015. The data provided are summarised to the postcode district level. Further information on this research is available in a research publication by James Aegerter, David Fouracre & Graham C. Smith, discussing the structure and density of pet cat and dog populations across Great Britain.

ABOUT DATASET

The 203 thermal infrared images comprising the Roboflow Thermal canines and People dataset were captured at varying distances from canines and people in a park and in the vicinity of a residence. Certain images are purposefully devoid of annotations due to the absence of faces or canines (refer to the Dataset Health Check for further details). There were both portrait and landscape photographs taken. (Auto-orientation ensures that annotations are aligned irrespective of the image orientation.)

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F15584165%2Fd6cb75db00e7a30d459b527493b14218%2FIMG_0099_jpg.rf.9546da350644619ff3164ea81d259120.jpg?generation=1701316984463514&alt=media" alt="">

Usage

The utilisation of thermal images is multifaceted, encompassing tasks such as machine performance monitoring, low-light visibility, and the enhancement of standard RGB scenarios with an additional dimension. Infrared imaging has applications in outdoor recreation, wildlife detection, and security.

Content From:

• https://twitter.com/josephofiowa
• https://public.roboflow.com/

People, Dogs And Monkeys

## Overview

People, Dogs And Monkeys is a dataset for object detection tasks - it contains People annotations for 3,248 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).

Introduction: Diminishing cognitive and physical functions, worsening psychological symptoms, and increased mortality risk and morbidity typically accompany aging. The aging population's health needs will continue to increase as the proportion of the population aged > 50 years increases. Pet ownership (PO) has been linked to better health outcomes in older adults, particularly those with chronic conditions. Much of the evidence is weak. Little is known about PO patterns as people age or the contribution of PO to successful aging in community-dwelling older adults. This study examines PO patterns among healthy community-dwelling older adults and the relationship of PO to cognitive and physical functions and psychological status.Methods: Participants in the Baltimore Longitudinal Study of Aging (> 50 years old, N = 378) completed a battery of cognitive, physical function, and psychological tests, as well as a PO questionnaire. Descriptive and non-parametric or general/generalized linear model analyses were conducted for separate outcomes.Results: Most participants (82%) had kept pets and 24% have pets: 14% dogs, 12% cats, 3% other pets. The most frequent reasons for having pets included enjoyment (80%) and companionship (66%). Most owners had kept the pet they had the longest for over 10 years (70%). PO was lower in older decades (p < 0.001). Pet owners were more likely to live in single-family homes and reside with others (p = 0.001) than non-owners. Controlling for age, PO was associated independently with better cognitive function (verbal leaning/memory p = 0.041), dog ownership predicted better physical function (daily energy expenditure, p = 0.018), and cat ownership predicted better cognitive functioning (verbal learning/memory, p = 0.035). Many older adults who did not own pets (37%) had regular contact with pets, which was also related to health outcomes.Conclusion: PO is lower at older ages, which mirrors the general pattern of poorer cognitive and physical function, and psychological status at older ages. PO and regular contact with pets (including PO) are associated with better cognitive status compared with those who did not own pets or had no regular contact with pets independent of age. Dog ownership was related to better physical function. Longitudinal analysis is required to evaluate the association of PO and/or regular contact with maintenance of health status over time.

In this competition, you'll write an algorithm to classify whether images contain either a dog or a cat. This is easy for humans, dogs, and cats. Your computer will find it a bit more difficult.

https://www.ethosvet.com/wp-content/uploads/cat-dog-625x375.png" alt="">

The Asirra data set

Web services are often protected with a challenge that's supposed to be easy for people to solve, but difficult for computers. Such a challenge is often called a CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) or HIP (Human Interactive Proof). HIPs are used for many purposes, such as to reduce email and blog spam and prevent brute-force attacks on web site passwords.

Asirra (Animal Species Image Recognition for Restricting Access) is a HIP that works by asking users to identify photographs of cats and dogs. This task is difficult for computers, but studies have shown that people can accomplish it quickly and accurately. Many even think it's fun! Here is an example of the Asirra interface:

Asirra is unique because of its partnership with Petfinder.com, the world's largest site devoted to finding homes for homeless pets. They've provided Microsoft Research with over three million images of cats and dogs, manually classified by people at thousands of animal shelters across the United States. Kaggle is fortunate to offer a subset of this data for fun and research. Image recognition attacks

While random guessing is the easiest form of attack, various forms of image recognition can allow an attacker to make guesses that are better than random. There is enormous diversity in the photo database (a wide variety of backgrounds, angles, poses, lighting, etc.), making accurate automatic classification difficult. In an informal poll conducted many years ago, computer vision experts posited that a classifier with better than 60% accuracy would be difficult without a major advance in the state of the art. For reference, a 60% classifier improves the guessing probability of a 12-image HIP from 1/4096 to 1/459. State of the art

The current literature suggests machine classifiers can score above 80% accuracy on this task [1]. Therfore, Asirra is no longer considered safe from attack. We have created this contest to benchmark the latest computer vision and deep learning approaches to this problem. Can you crack the CAPTCHA? Can you improve the state of the art? Can you create lasting peace between cats and dogs?

Submission Format

Your submission should have a header. For each image in the test set, predict a label for its id (1 = dog, 0 = cat):

id,label 1,0 2,0 3,0 etc...

Cat&dog&people

## Overview

Cat&dog&people is a dataset for object detection tasks - it contains Cats annotations for 612 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).

The present dataset is based on a questionnaire which is also part of this package. The enclose questionnaire includes identifiable and relevant variables names (yellow highlighted).

Participants were recruited by Norstat, a European-based survey company, with the aim of gaining a representative sample of Austrian, Danish and UK citizens, including pet owners. The survey company administers and hosts online panels comprising citizens from many European countries. We aimed for a sample that is representative in terms of age, gender, and region. Therefore, a stratified sampling principle was set up where individuals within each stratum were randomly invited to participate. The invitations were issued through e-mail that contained a link to the online questionnaire. Data was collected from 11-25th of March 2022 in Austria, from 11-24th of March 2022 in Denmark and from 8-23rd of March 2022 in the UK. The invitation provided information about the background of the study, the participating universities, ethical approval, estimated time for questionnaire completion and further, participants were informed that the completion of the questionnaire was voluntary and anonymous, and that they could exit the survey at any point. Before participants were directed to the survey, they ensured informed consent by confirming that they are over 17 years old, and consent to participate in this survey.

Besides the questionnaire the dataset includes a csv and an Excel file consisting of the data that is used in the ms. and an rtf and a pdf file with data variable names/labels, and value labels.

Over the last century, dogs have been increasingly used to detect rare and elusive species or traces of them. The use of wildlife detection dogs (WDD) is particularly well established in North America, Europe and Oceania, and projects deploying them have increased worldwide. However, if they are to make a significant contribution to conservation and management, their strengths, abilities, and limitations should be fully identified. We reviewed the use of WDD with particular focus on the breeds used in different countries and for various targets, as well as their overall performance compared to other methods, by developing and analysing a database of 1220 publications, including 916 scientific ones, covering 2464 individual cases - most of them (1840) scientific. With the worldwide increase in the use of WDD, associated tasks have changed and become much more diverse. Since 1930, reports exist for 62 countries and 407 animal, 42 plant, 26 fungi and 6 bacteria species. Altogether, 108 FCI-classified and 20 non-FCI-classified breeds have worked as WDD. While certain breeds have been preferred on different continents and for specific tasks and targets, they were not generally better suited for detection tasks than others. Overall, WDD usually worked more effectively than other monitoring methods. For each species group, regardless of breed, detection dogs were better than other methods in 88.71% of all cases and only worse in 0.98%. It was only for arthropods that Pinshers and Schnauzers performed worse than other breeds. For mono- and dicotyledons, detection dogs did less often outperform other methods. Although every breed can be trained as a WDD, choosing the most suitable dog for the task and target may speed up training and increase the chance of success. Albeit selection of the most appropriate WDD is important, excellent training, knowledge about the target density and suitability, and a proper study design all appeared to have the highest impact on performance. Moreover, an appropriate area, habitat and weather are crucial for detection dog work. When these factors are taken into consideration, WDD can be an outstanding monitoring method.

Methods We systematically searched for any publication using the following search terms in Google Scholar and ISI Web of Knowledge: wildlife detect* dog, species detect* dog, scat detect* dog, [species] + detect* dog, [author] + detect* dog, [country] + detect* dog, conservation (detect*) dog, predator (detect*) dog, protected species (detect*) dog, den detect* dog, roost detect* dog, plant detect* dog, canine detection, and tracking dog. We traced any potentially relevant cited publication and only included those in our review that we could check ourselves. We also collected publications if we got to know them otherwise and reviewed existing literature lists and compilations (Grimm-Seyfarth et al. 2021, Appendix S1.1). We focused mainly on scientific literature, including scientific papers, dissertations, and project reports. However, wildlife detection dogs were frequently used for conservation or management purposes without a scientific research project behind them. For a more comprehensive overview of their deployment and performance, we included popular science or newspaper articles when no scientific publication about the project was found. In addition, we used social media platforms to obtain many articles from different countries (Grimm-Seyfarth et al. 2021, Appendix S1.1). In order to avoid multiple citations of the same study for which publications from different sources have been published, we compared each new entry with the entries in the database and preferably included scientific publications, followed by books, popular science and newspaper articles.

We compiled the data in a relational database (Microsoft Access 2013) consisting of five basic tables: literature, dog breeds, target species, target types and countries. We classified dog breeds into the ten FCI classification groups and breeds not listed as “not classified”. We assigned mixed breeds to a main or first-mentioned breed or to the category “Mix” when they could not be assigned to a specific breed. We classified target species according to their Latin and English names, genus, family, order, class, phylum and kingdom, adding subspecies names if provided. If the dog detected species groups without further specification (e.g., bat or bird carcasses, rodents, weed), we retained this group only. Taxonomic changes due to splitting of taxa into several species were only made if the allocation to the new species was obvious from the geographic information provided or had already been done by other authors. We divided potential target types into: living or dead individuals; nests, dens, clutches, coveys, roosts; scat, urine, saliva, glandular secretion; spores, eggs; larvae; hair, feathers, pellets, shed skin; and different combinations thereof. Lastly, we classified countries according to the (sub-) continent into North, Central and South America, Europe, Asia, Africa, and Oceania, assigning Russia and Turkey to “Eurasia”. Furthermore, we assigned Australia, New Zealand, and all oceanic islands (including subantarctic islands) to “Oceania” and made no differentiation to Zealandia.

In a main table, we then assigned each breed-target species-country association per reference as a single “case”. We marked pure-breed dogs and added a second breed for mixed breeds (if provided), as well as the number of dogs per breed and reference (if not mentioned directly, “1” for mentioning “dog” and “2” for mentioning “dogs”). We also added specifications to the country (e.g. Islands). If available, we extracted results of the wildlife detection dog performance compared to other monitoring methods. We classified the performance into four categories: dogs were (i) better; (ii) equal; or (iii) worse than other methods tested; or (iv) mixed results. The factor in comparison was study-specific and could include speed per area or transect, area size, sample size, quality, detectability, specificity, sensitivity, accuracy, or precision. We relied on those conservative measures since different monitoring methods can hardly be compared otherwise. The category “mixed results” was given when the dogs were better at some factors but worse at others, or when the performance depended upon season, year, site, or dog. Since we designed the database as a relational database, IDs among the five basic tables and the main table were linked together for quick searches and queries.

About This Dataset

The Roboflow Thermal Dogs and People dataset is a collection of 203 thermal infrared images captured at various distances from people and dogs in a park and near a home. Some images are deliberately unannotated as they do not contain a person or dog (see the Dataset Health Check for more). Images were captured both portrait and landscape. (Roboflow auto-orient assures the annotations align regardless of the image orientation.)

Thermal images were captured using the Seek Compact XR Extra Range Thermal Imaging Camera for iPhone. The selected color palette is Spectra.

Example

This is an example image and annotation from the dataset: https://i.imgur.com/h9vhrqB.png" alt="Man and Dog">

Usage

Thermal images have a wide array of applications: monitoring machine performance, seeing in low light conditions, and adding another dimension to standard RGB scenarios. Infrared imaging is useful in security, wildlife detection,and hunting / outdoors recreation.

This dataset serves as a way to experiment with infrared images in Roboflow. (Or, you could build your own night time pet finder!)

Collecting Custom Data

Roboflow is happy to improve your operations with infrared imaging and computer vision. Services range from data collection to building automated monitoring systems leveraging computer vision. Reach out for more.

About Roboflow

Roboflow makes managing, preprocessing, augmenting, and versioning datasets for computer vision seamless. :fa-spacer: Developers reduce 50% of their boilerplate code when using Roboflow's workflow, save training time, and increase model reproducibility. :fa-spacer: