The IMPTOX project has received funding from the EU's H2020 framework programme for research and innovation under grant agreement n. 965173. Imptox is part of the European MNP cluster on human health.

More information about the project here.

Description: This repository includes the trained weights and a custom COCO-formatted dataset used for developing and testing a Faster R-CNN R_50_FPN_3x object detector, specifically designed to identify particles in micro-FTIR filter images.

Contents:

Weights File (neuralNetWeights_V3.pth):

Format: .pth

Description: This file contains the trained weights for a Faster R-CNN model with a ResNet-50 backbone and a Feature Pyramid Network (FPN), trained for 3x schedule. These weights are specifically tuned for detecting particles in micro-FTIR filter images.

Custom COCO Dataset (uFTIR_curated_square.v5-uftir_curated_square_2024-03-14.coco-segmentation.zip):

Format: .zip

Description: This zip archive contains a custom COCO-formatted dataset, including JPEG images and their corresponding annotation file. The dataset consists of images of micro-FTIR filters with annotated particles.

Contents:

Images: JPEG format images of micro-FTIR filters.

Annotations: A JSON file in COCO format providing detailed annotations of the particles in the images.

Management: The dataset can be managed and manipulated using the Pycocotools library, facilitating easy integration with existing COCO tools and workflows.

Applications: The provided weights and dataset are intended for researchers and practitioners in the field of microscopy and particle detection. The dataset and model can be used for further training, validation, and fine-tuning of object detection models in similar domains.

Usage Notes:

The neuralNetWeights_V3.pth file should be loaded into a PyTorch model compatible with the Faster R-CNN architecture, such as Detectron2.

The contents of uFTIR_curated_square.v5-uftir_curated_square_2024-03-14.coco-segmentation.zip should be extracted and can be used with any COCO-compatible object detection framework for training and evaluation purposes.

Code can be found on the related Github repository.

SpeechCoco

Introduction

Our corpus is an extension of the MS COCO image recognition and captioning dataset. MS COCO comprises images paired with a set of five captions. Yet, it does not include any speech. Therefore, we used Voxygen's text-to-speech system to synthesise the available captions.

The addition of speech as a new modality enables MSCOCO to be used for researches in the field of language acquisition, unsupervised term discovery, keyword spotting, or semantic embedding using speech and vision.

Our corpus is licensed under a Creative Commons Attribution 4.0 License.

Data Set

• This corpus contains 616,767 spoken captions from MSCOCO's val2014 and train2014 subsets (respectively 414,113 for train2014 and 202,654 for val2014).

• We used 8 different voices. 4 of them have a British accent (Paul, Bronwen, Judith, and Elizabeth) and the 4 others have an American accent (Phil, Bruce, Amanda, Jenny).

• In order to make the captions sound more natural, we used SOX tempo command, enabling us to change the speed without changing the pitch. 1/3 of the captions are 10% slower than the original pace, 1/3 are 10% faster. The last third of the captions was kept untouched.

• We also modified approximately 30% of the original captions and added disfluencies such as "um", "uh", "er" so that the captions would sound more natural.

• Each WAV file is paired with a JSON file containing various information: timecode of each word in the caption, name of the speaker, name of the WAV file, etc. The JSON files have the following data structure:

{
  "duration": float,
  "speaker": string,
  "synthesisedCaption": string,
  "timecode": list,
  "speed": float,
  "wavFilename": string,
  "captionID": int,
  "imgID": int,
  "disfluency": list
}

• On average, each caption comprises 10.79 tokens, disfluencies included. The WAV files are on average 3.52 seconds long.

Repository

The repository is organized as follows:

• CORPUS-MSCOCO (~75GB once decompressed)

  • train2014/ : folder contains 413,915 captions

    • json/

    • wav/

    • translations/

      • train_en_ja.txt

      • train_translate.sqlite3

    • train_2014.sqlite3

  • val2014/ : folder contains 202,520 captions

    • json/

    • wav/

    • translations/

      • train_en_ja.txt

      • train_translate.sqlite3

    • val_2014.sqlite3

  • speechcoco_API/

    • speechcoco/

      • _init_.py

      • speechcoco.py

    • setup.py

Filenames

.wav files contain the spoken version of a caption

.json files contain all the metadata of a given WAV file

.sqlite3 files are SQLite databases containing all the information contained in the JSON files

We adopted the following naming convention for both the WAV and JSON files:

imageID_captionID_Speaker_DisfluencyPosition_Speed[.wav/.json]

Script

We created a script called speechcoco.py in order to handle the metadata and allow the user to easily find captions according to specific filters. The script uses the *.db files.

Features:

• Aggregate all the information in the JSON files into a single SQLite database

• Find captions according to specific filters (name, gender and nationality of the speaker, disfluency position, speed, duration, and words in the caption). The script automatically builds the SQLite query. The user can also provide his own SQLite query.

The following Python code returns all the captions spoken by a male with an American accent for which the speed was slowed down by 10% and that contain "keys" at any position

# create SpeechCoco object
db = SpeechCoco(train_2014.sqlite3, train_translate.sqlite3, verbose=True)

# filter captions (returns Caption Objects)
captions = db.filterCaptions(gender="Male", nationality="US", speed=0.9, text='%keys%')
for caption in captions:
  print('
{}\t{}\t{}\t{}\t{}\t{}\t\t{}'.format(caption.imageID,
                         caption.captionID,
                         caption.speaker.name,
                         caption.speaker.nationality,
                         caption.speed,
                         caption.filename,
                         caption.text))

...
298817   26763  Phil  0.9   298817_26763_Phil_None_0-9.wav     A group of turkeys with bushes in the background.
108505   147972 Phil  0.9   108505_147972_Phil_Middle_0-9.wav        Person using a, um, slider cell phone with blue backlit keys.
258289   154380 Bruce  0.9   258289_154380_Bruce_None_0-9.wav        Some donkeys and sheep are in their green pens .
545312   201303 Phil  0.9   545312_201303_Phil_None_0-9.wav     A man walking next to a couple of donkeys.
...

• Find all the captions belonging to a specific image

captions = db.getImgCaptions(298817)
for caption in captions:
  print('
{}'.format(caption.text))

Birds wondering through grassy ground next to bushes.
A flock of turkeys are making their way up a hill.
Um, ah. Two wild turkeys in a field walking around.
Four wild turkeys and some bushes trees and weeds.
A group of turkeys with bushes in the background.

• Parse the timecodes and have them structured

input:

...
[1926.3068, "SYL", ""],
[1926.3068, "SEPR", " "],
[1926.3068, "WORD", "white"],
[1926.3068, "PHO", "w"],
[2050.7955, "PHO", "ai"],
[2144.6591, "PHO", "t"],
[2179.3182, "SYL", ""],
[2179.3182, "SEPR", " "]
...

output:

print(caption.timecode.parse())

...
{
'begin': 1926.3068,
'end': 2179.3182,
'syllable': [{'begin': 1926.3068,
       'end': 2179.3182,
       'phoneme': [{'begin': 1926.3068,
              'end': 2050.7955,
              'value': 'w'},
             {'begin': 2050.7955,
              'end': 2144.6591,
              'value': 'ai'},
             {'begin': 2144.6591,
              'end': 2179.3182,
              'value': 't'}],
       'value': 'wait'}],
'value': 'white'
},
...

• Convert the timecodes to Praat TextGrid files

caption.timecode.toTextgrid(outputDir, level=3)

• Get the words, syllables and phonemes between n seconds/milliseconds

The following Python code returns all the words between 0.2 and 0.6 seconds for which at least 50% of the word's total length is within the specified interval

pprint(caption.getWords(0.20, 0.60, seconds=True, level=1, olapthr=50))

...
404537   827239 Bruce  US   0.9   404537_827239_Bruce_None_0-9.wav        Eyeglasses, a cellphone, some keys and other pocket items are all laid out on the cloth. .
[
  {
    'begin': 0.0,
    'end': 0.7202778,
    'overlapPercentage': 55.53412863758955,
    'word': 'eyeglasses'
  }
]
 ...

• Get the translations of the selected captions

As for now, only japanese translations are available. We also used Kytea to tokenize and tag the captions translated with Google Translate

captions = db.getImgCaptions(298817)
for caption in captions:
  print('
{}'.format(caption.text))

  # Get translations and POS
  print('\tja_google: {}'.format(db.getTranslation(caption.captionID, "ja_google")))
  print('\t\tja_google_tokens: {}'.format(db.getTokens(caption.captionID, "ja_google")))
  print('\t\tja_google_pos: {}'.format(db.getPOS(caption.captionID, "ja_google")))
  print('\tja_excite: {}'.format(db.getTranslation(caption.captionID, "ja_excite")))

  Birds wondering through grassy ground next to bushes.
  ja_google: 鳥は茂みの下に茂った地面を抱えています。
    ja_google_tokens: 鳥 は 茂み の 下 に 茂 っ た 地面 を 抱え て い ま す 。
    ja_google_pos: 鳥/名詞/とり は/助詞/は 茂み/名詞/しげみ の/助詞/の 下/名詞/した に/助詞/に

IoU and HD of UNets family when the optimizer is set as Adam. lr: learning rate. The best-performing results are highlighted using bold font.

numoverseg and nummiss of three BTS models. The best-performing results are highlighted using bold font. lr: learning rate.

Automatic image captioning is the task of producing a natural-language utterance (usually a sentence) that correctly reflects the visual content of an image. Up to this point, the resource most used for this task was the MS-COCO dataset, containing around 120,000 images and 5-way image-caption annotations (produced by paid annotators). Google's Conceptual Captions dataset has more than 3 million images, paired with natural-language captions. In contrast with the curated style of the MS-COCO images, Conceptual Captions images and their raw descriptions are harvested from the web, and therefore represent a wider variety of styles. The raw descriptions are harvested from the Alt-text HTML attribute associated with web images. The authors developed an automatic pipeline that extracts, filters, and transforms candidate image/caption pairs, with the goal of achieving a balance of cleanliness, informativeness, fluency, and learnability of the resulting captions.

LAION COCO with aesthetic score and watermark score

This dataset contains 10% samples of the LAION-COCO dataset filtered by some text rules (remove url, special tokens, etc.), and image rules (image size > 384x384, aesthetic score>4.75 and watermark probability<0.5). There are total 8,563,753 data instances in this dataset. And the corresponding aesthetic score and watermark score are also included. Noted: watermark score in the table means the probability of the existence of the… See the full description on the dataset page: https://huggingface.co/datasets/guangyil/laion-coco-aesthetic.

The global carbon block filter market, valued at $1671.2 million in 2025, is projected to experience steady growth, driven by increasing demand for clean and safe drinking water. The compound annual growth rate (CAGR) of 3.4% from 2025 to 2033 indicates a consistent expansion, fueled by rising concerns about waterborne contaminants and the increasing adoption of point-of-use (POU) and point-of-entry (POE) water filtration systems in both residential and commercial settings. Key drivers include the growing awareness of water contamination issues, stricter government regulations regarding water quality, and the rising disposable incomes in developing economies leading to increased demand for improved water filtration solutions. Furthermore, advancements in carbon block filter technology, such as the development of more efficient and long-lasting filters, are contributing to market growth. The market is segmented by type (e.g., granular activated carbon, carbon block), application (residential, commercial), and region. Competitive landscape analysis reveals key players such as Marmon, Multipure, and others actively innovating and expanding their product portfolios to cater to evolving consumer preferences. Despite the positive growth outlook, certain challenges exist. Fluctuations in raw material prices, especially activated carbon, can impact production costs and profitability. Furthermore, the presence of numerous smaller players creates a competitive market environment. However, the increasing emphasis on sustainable water management practices and the growing adoption of water purification technologies are expected to offset these restraints and drive long-term market expansion. The market is geographically diverse, with North America and Europe currently holding significant market shares. However, developing economies in Asia-Pacific and other regions are expected to witness rapid growth due to increasing urbanization and rising awareness of water quality issues. The forecast period of 2025-2033 promises continued expansion, driven by technological advancements and a growing global focus on water safety and purity.

IoU and HD of UNets family when the optimizer is set as Adagrad. The best-performing results are highlighted using bold font.

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}
 }

DF2023 is a dataset for image forgery detection and localization. The training and validation datasets contain 1,000,000/5,000 manipulated images (and the ground truth masks).

The DF2023 training dataset comprises:

• 100K forged images produced by removal (inpainting) operations
• 200K images produced by enhancement modifications
• 300K copy-move manipulated images and
• 400K spliced images

=== 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.

=== Terms of Use / Licence ===

The DF2023 dataset is based on the MS COCO dataset. Therefore, rules for using the images form MS COCO apply also for DF2023:

Images

The COCO Consortium does not own the copyright of the images. Use of the images must abide by the Flickr Terms of Use. The users of the images accept full responsibility for the use of the dataset, including but not limited to the use of any copies of copyrighted images that they may create from the dataset.

Here are a few use cases for this project:

1. Human Presence Detection: This computer vision model can be incorporated into security systems and smart home devices to identify the presence of humans in an area, allowing for customized responses, room automation, and improved safety.

2. Crowd Size Estimation: The "human dataset v1" can be used by event organizers or city planners to estimate the size of gatherings or crowds at public events, helping them better allocate resources and manage these events more efficiently.

3. Surveillance and Security Enhancement: The model can be integrated into video surveillance systems to more accurately identify humans, helping to filter out false alarms caused by animals and other non-human entities.

4. Collaborative Robotics: Robots equipped with this computer vision model can more easily identify and differentiate humans from their surroundings, allowing them to more effectively collaborate with people in shared spaces while ensuring human safety.

5. Smart Advertising: The "human dataset v1" can be utilized by digital signage and advertising systems to detect and count the number of human viewers, enabling targeted advertising and measuring the effectiveness of marketing campaigns.

Overview

The original Udacity Self Driving Car Dataset is missing labels for thousands of pedestrians, bikers, cars, and traffic lights. This will result in poor model performance. When used in the context of self driving cars, this could even lead to human fatalities.

We re-labeled the dataset to correct errors and omissions. We have provided convenient downloads in many formats including VOC XML, COCO JSON, Tensorflow Object Detection TFRecords, and more.

Some examples of labels missing from the original dataset: https://i.imgur.com/A5J3qSt.jpg" alt="Examples of Missing Labels">

Stats

The dataset contains 97,942 labels across 11 classes and 15,000 images. There are 1,720 null examples (images with no labels).

All images are 1920x1200 (download size ~3.1 GB). We have also provided a version downsampled to 512x512 (download size ~580 MB) that is suitable for most common machine learning models (including YOLO v3, Mask R-CNN, SSD, and mobilenet).

Annotations have been hand-checked for accuracy by Roboflow.

https://i.imgur.com/bOFkueI.pnghttps://" alt="Class Balance">

Annotation Distribution: https://i.imgur.com/NwcrQKK.png" alt="Annotation Heatmap">

Use Cases

Udacity is building an open source self driving car! You might also try using this dataset to do person-detection and tracking.

Using this Dataset

Our updates to the dataset are released under the MIT License (the same license as the original annotations and images).

Note: the dataset contains many duplicated bounding boxes for the same subject which we have not corrected. You will probably want to filter them by taking the IOU for classes that are 100% overlapping or it could affect your model performance (expecially in stoplight detection which seems to suffer from an especially severe case of duplicated bounding boxes).

About Roboflow

Roboflow makes managing, preprocessing, augmenting, and versioning datasets for computer vision seamless.

Developers reduce 50% of their boilerplate code when using Roboflow's workflow, save training time, and increase model reproducibility. :fa-spacer:

Fusion results with UNet2+ (optimizer = Adagrad, learning rate = 0.0001). The best-performing results are highlighted using bold font.

This dataset is curated from subsets of public datasets such as MS COCO, LAION-ART, and SBU Captions. It specifically filters for samples featuring dog-related content. The goal of this dataset is to support text-to-image generation models focused on dog objects and related scenes. https://github.com/rom1504/img2dataset/blob/main/dataset_examples/SBUcaptions.md https://github.com/rom1504/img2dataset/blob/main/dataset_examples/mscoco.md… See the full description on the dataset page: https://huggingface.co/datasets/MikdadMrhij/Dogs-images-text-pair.

Introduction

This dataset supports Ye et al. 2024 Nature Communications. Please cite this dataset and paper if you use this resource. Please also see Ye et al. 2024 for the full DataSheet that accompanies this download, including the meta data for how to use this data is you want to compare model results on benchmark tasks. Below is just a summary. Also see the dataset licensing below.

Training Data

It consists of being trained together on the following datasets:

• AwA-Pose Quadruped dataset, see full details at (1).
• AnimalPose See full details at (2).
• AcinoSet See full details at (3).
• Horse-30 Horse-30 dataset, benchmark task is called Horse-10; See full details at (4).
• StanfordDogs See full details at (5, 6).
• AP-10K See full details at (7).
• iRodent We utilized the iNaturalist API functions for scraping observations with the taxon ID of Suborder Myomorpha (8). The functions allowed us to filter the large amount of observations down to the ones with photos under the CC BY-NC creative license. The most common types of rodents from the collected observations are Muskrat (Ondatra zibethicus), Brown Rat (Rattus norvegicus), House Mouse (Mus musculus), Black Rat (Rattus rattus), Hispid Cotton Rat (Sigmodon hispidus), Meadow Vole (Microtus pennsylvanicus), Bank Vole (Clethrionomys glareolus), Deer Mouse (Peromyscus maniculatus), White-footed Mouse (Peromyscus leucopus), Striped Field Mouse (Apodemus agrarius). We then generated segmentation masks over target animals in the data by processing the media through an algorithm we designed that uses a Mask Region Based Convolutional Neural Networks(Mask R-CNN) (9) model with a ResNet-50-FPN backbone (10), pretrained on the COCO datasets (11). The processed 443 images were then manually labeled with both pose annotations and segmentation masks. iRodent data is banked at https://zenodo.org/record/8250392.
• APT-36K See full details at (12).

https://images.squarespace-cdn.com/content/v1/57f6d51c9f74566f55ecf271/1690988780004-AG00N6OU1R21MZ0AU9RE/modelcard-SAQ.png?format=1500w" target="_blank" rel="noopener">Here is an image with a keypoint guide.

Ethical Considerations

• No experimental data was collected for this model; all datasets used are cited above.

Caveats and Recommendations

• Please note that each dataest was labeled by separate labs & separate individuals, therefore while we map names to a unified pose vocabulary, there will be annotator bias in keypoint placement (See Ye et al. 2024 for our Supplementary Note on annotator bias). You will also note the dataset is highly diverse across species, but collectively has more representation of domesticated animals like dogs, cats, horses, and cattle. We recommend if performance of a model trained on this data is not as good as you need it to be, first try video adaptation (see Ye et al. 2024), or fine-tune the weights with your own labeling.

License

Modified MIT.

Copyright 2023-present by Mackenzie Mathis, Shaokai Ye, and contributors.

Permission is hereby granted to you (hereafter "LICENSEE") a fully-paid, non-exclusive,
and non-transferable license for academic, non-commercial purposes only (hereafter “LICENSE”)
to use the "DATASET" subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial
portions of the Software:

This data or resulting software may not be used to harm any animal deliberately.

LICENSEE acknowledges that the DATASET is a research tool.
THE DATASET IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING
BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE DATASET OR THE USE OR OTHER DEALINGS IN THE DATASET.

If this license is not appropriate for your application, please contact Prof. Mackenzie W. Mathis
(mackenzie@post.harvard.edu) for a commercial use license.

Please cite Ye et al if you use this DATASET in your work.

References

1. Prianka Banik, Lin Li, and Xishuang Dong. A novel dataset for keypoint detection of quadruped animals from images. ArXiv, abs/2108.13958, 2021
2. Jinkun Cao, Hongyang Tang, Haoshu Fang, Xiaoyong Shen, Cewu Lu, and Yu-Wing Tai. Cross-domain adaptation for animal pose estimation. 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pages 9497–9506, 2019.
3. Daniel Joska, Liam Clark, Naoya Muramatsu, Ricardo Jericevich, Fred Nicolls, Alexander Mathis, Mackenzie W. Mathis, and Amir Patel. Acinoset: A 3d pose estimation dataset and baseline models for cheetahs in the wild. 2021 IEEE International Conference on Robotics and Automation (ICRA), pages 13901–13908, 2021.
4. Alexander Mathis, Thomas Biasi, Steffen Schneider, Mert Yuksekgonul, Byron Rogers, Matthias Bethge, and Mackenzie W Mathis. Pretraining boosts out-of-domain robustness for pose estimation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 1859–1868, 2021.
5. Aditya Khosla, Nityananda Jayadevaprakash, Bangpeng Yao, and Li Fei-Fei. Novel dataset for fine-grained image categorization. In First Workshop on Fine-Grained Visual Categorization, IEEE Conference on Computer Vision and Pattern Recognition, Colorado Springs, CO, June 2011.
6. Benjamin Biggs, Thomas Roddick, Andrew Fitzgibbon, and Roberto Cipolla. Creatures great and smal: Recovering the shape and motion of animals from video. In Asian Conference on Computer Vision, pages 3–19. Springer, 2018.
7. Hang Yu, Yufei Xu, Jing Zhang, Wei Zhao, Ziyu Guan, and Dacheng Tao. Ap-10k: A benchmark for animal pose estimation in the wild. In Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2), 2021.
8. iNaturalist. OGBIF Occurrence Download. https://doi.org/10.15468/dl.p7nbxt. iNaturalist, July 2020
9. Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. Mask r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 2961–2969, 2017.
10. Tsung-Yi Lin, Piotr Dollár, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature pyramid networks for object detection, 2016.
11. Tsung-Yi Lin, Michael Maire, Serge J. Belongie, Lubomir D. Bourdev, Ross B. Girshick, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll’ar, and C. Lawrence Zitnick. Microsoft COCO: common objects in context. CoRR, abs/1405.0312, 2014
12. Yuxiang Yang, Junjie Yang, Yufei Xu, Jing Zhang, Long Lan, and Dacheng Tao. Apt-36k: A large-scale benchmark for animal pose estimation and tracking. Advances in Neural Information Processing Systems, 35:17301–17313, 2022

Versioning Note:

- V2 includes fixes to Stanford Dog data; it affected less than 1% of the data.

This dataset, FMPD (Freshwater Microscopy Phytoplankton Dataset), is released for non-comercial academic or research purposes only, subject to attribution through citation of the following papers

• Figueroa, J. Rouco, J. Novo, "Phytoplankton detection and recognition in freshwater digital microscopy images using deep learning object detectors", Heliyon, 2023.

• D. Rivas-Villar, J. Rouco, M. G. Penedo, R. Carballeira, J. Novo, "Fully automatic detection and classification of phytoplankton specimens in digital microscopy images", Computer Methods and Programs in Biomedicine, 200, 105923, 2021

Please also consider the citation of any of the other related papers from the dataset authors.

Data:

The FMPD dataset is a set of multi-specimen microscopy images of freshwater phytoplankton. These images have been captured with fixed settings, equal for each image, including illumination, focal point and magnification. The dataset contains 293 images from water sampled at lake of Doniños (Ferrol, Galicia, Spain) (UTM 555593 X, 4815672 Y; Datum ETRS89) on multiple visits throughout the year. This ensures seasonal representability.

The phytoplankton sample was concentrated by filtering volume of 0.5 L through GF/F glass fiber filters and was then resuspended in 50 mL. Phytoplankton samples were preserved using 5% (v/v) glutaraldehyde, because it is efficient at preserving both cellular structures and pigment. The fixed sample was stored in the dark at constant temperature (10 oC) until analysis. The phytoplankton sample was homogenised for 2 min prior to microscopic examination. In addition, the sample was subjected to vacuum for one minute to break the vacuoles of some cyanobacterial taxa and prevent them from floating. Aliquots of the phytoplankton sample with a total volume of 1 mL were examined under light microscopy using a Nikon Eclipse E600 equipped with an E-Plan 10× objective (N.A. 0.25). Light microscopy images were taken with an AxioCam ICc5 Zeiss digital camera, maintaining the same illumination and focus throughout the image acquisition process and following regular transects until the entire surface of the sample was covered.

The dataset contains 293 multi-specimen phytoplankton images. As mentioned, these images have fixed magnification, illumination and focal point. The produced images are saved in .tif format with a size of 2080x1540 pixels and are located in the dataset folder. The ground truth consists of bounding boxes that enclose the phytoplankton specimens, with an associated label identifying the species. Currently, this dataset has tags for:

• Non-phytoplankton: particles, debris, zooplankton or any other object that could be mistaken as phytoplankton- Woronichinia naegeliana: Toxin-producing cyanobacteria- Anabaena Spiroides: Toxin-producing cyanobacteria- Dinobryon Sp.: Harmless but challenging as it can both appear solitary or in colonies- Other-phytoplankton: Other phytoplankton species.

Annotations are provided in a .json file in the format typically used by the coco dataset, in the annotations.json file.

Holdout train-test splits, as well as k-fold cross-validation splits, are provided in the splits folder, available in .json format. These splits correspond to those used in the previously mentioned papers to be cited, facilitating straightforward comparisons. Additionally, the annotations for each subset are included in separate files within the same folder for ease of use. It should be noted that the annotations.json contains all of these subsets of annotations.

In 2022, brand awareness of Kellogg's Coco Pops was ** percent among survey respondents. When asked how many UK consumers ate Coco Pops in the past 12 months, just over a third of respondents stated that they had.

Want more brand data?Explore more Statista Brand Profiles. If you want to compare brands, do deep-dives by survey items of your choice, filter by total online population or users of a certain brand, or drill down on your very own hand-tailored target groups, our Statista Brand Profiler has you covered.

There are 3 child zip files included in this data release. 01_Codebase.zip contains a codebase for using deep learning to filter images based on the probability of any bird occurrence. It includes instructions and files necessary for training, validating, and testing a machine learning detection algorithm. 02_Imagery.zip contains imagery that were collected using a Partenavia P68 fixed-wing airplane using a PhaseOne iXU-R 180 forward motion compensating 80-megapixel digital frame camera with a 70 mm Rodenstock lens. The imagery were cropped into smaller patches of 720x720 pixels for training and 1440x1440 pixels for validation and test datasets. These data were collected for developing machine learning algorithms for the detection and classification of avian targets in aerial imagery. These data can be paired with annotation values to train and evaluate object detection and classification models. 03_Annotations.zip contains a collection of bounding boxes around avian targets in aerial imagery formatted as COCO JSON file. The data are nested under evaluation and test folders and contain both ground truth targets and predicted targets.These data were collected for two main functions. The ground truth avian targets were manually annotated and can be used to train avian detection algorithms using machine learning methods. The predicted targets can be used to evaluate model performance while referencing the larger work associated with these data.

This dataset includes particulate carbon concentrations and isotopes collected in December, 2021. Suspended particles are collected at 4 different stations near Cocos Ridge, at two different size fractions using Mclane pumps. The two size fractions are large size fraction (LSF) that is >51 um, and small size fraction (SSF) that is 0.5 -- 51 um. Concentrations and stable carbon isotopes of particulate inorganic carbon (PIC) and total carbon (TC) are measured and reported. PIC content are measured by acidifying a subsample of the Glass Fiber Filter (GFF) and measuring total CO2 released using a Picarro Cavity Ring-down Spectroscopy. TC content are analyzed by burning a subsample of the GFF on Elemental Analyzer (EA). Samples were collected during SR2113 onboard Sally Ride, under the project "new approaches to study calcium carbonate dissolution on the sea floor and its impact on paleo-proxy interpretations", as a water-column side determination of particle compositions and carbonate dissolution. This data reveals changes in concentrations and stable carbon isotopes with water depth, and has implications for multiple biogeochemical processes associated with both the inorganic and the organic carbon within marine particles in the water column.

TreeAI - Advancing Tree Species Identification from Aerial Images with Deep Learning

• Register: Access to the data will be granted upon registering for the competition, see the registration form: https://form.ethz.ch/research/tree-ai-global-database/treeai-competition.html
• Request the dataset: Download the competition record after registration by requesting it. Enter your full name, purpose e.g. accept the TreeAI4Species data license, affiliation, and the country of affiliation in the request. This allows us to check whether you are already registered.
• Test dataset: Only the participants registered for the competition will receive the test dataset.
• Submit your DL models for evaluation by June 2025.
• Award: The best models win a prize.
• Publication: All participants in the competition who submit the required files for evaluation will be included in the subsequent publication.

License

Abstract Water is essential for life, important for the ecosystem and it is in great demand due to its scarcity. This study explored the reuse of the wastewater of the Water Treatment Plant in Gramame for agricultural purposes. A qualitative and quantitative investigation of the effluent was carried out through the characterization of its physical and chemical parameters, comparing the results to what is allowed by the current legislation, CONAMA Resolution 357/2005. After this process, the activated charcoal of coco-da-baia mesocarp, adsorbent material, was prepared and tested in a filter system in a column with a continuous flow and ascendant entrance, in which the kinetic effect was evaluated. This technique was evaluated by correlating the reduced values in the adsorption material, respecting the initial effluent concentration, obtaining a reduction of 50% in the hardness, 87.5% in chloride and 66.6% in acidity. These results verified the adequacy of the technique in potential hydrogenation (pH) and abrupt reduction of color and turbidity. This treatment is suggested to qualify the effluent for use in agricultural, safe for humans and the environment. The adsorbent substrate efficiency was verified by correlating it with the Thomas isothermal model.