Computer Vision Market size was valued at USD 11.22 billion in 2021 and is poised to grow from USD 12.01 billion in 2022 to USD 22.07 billion by 2030, growing at a CAGR of 7% in the forecast period (2023-2030).

Research Papers Collection

100 recent research papers from arXiv across 5 AI/data science topics.

  Dataset Overview

Total Papers: 100 (20 per topic) Source: arXiv API Format: PDF files Topics: Generative AI, Machine Learning, Statistics, Analytics, Computer Vision

  Topics & Queries

Generative_AI: cat:cs.LG AND (generative OR diffusion OR GAN OR transformer OR GPT) Machine_Learning: cat:cs.LG OR cat:stat.ML Statistics: cat:stat.TH OR cat:stat.ME OR cat:stat.AP… See the full description on the dataset page: https://huggingface.co/datasets/mahimaarora025/research_papers.

Context

The exact preprocessing steps used to construct the MNIST dataset have long been lost. This leaves us with no reliable way to associate its characters with the ID of the writer and little hope to recover the full MNIST testing set that had 60K images but was never released. The official MNIST testing set only contains 10K randomly sampled images and is often considered too small to provide meaningful confidence intervals.

The QMNIST dataset was generated from the original data found in the NIST Special Database 19 with the goal to match the MNIST preprocessing as closely as possible.

Content

The simplest way to use the QMNIST extended dataset is to download the unique file below (MNIST-120k). This pickle file has the same format as the standard MNIST data files but contains 120000 examples.

You can use the following lines of code to load the data: def unpickle(file): import pickle with open(file, 'rb') as fo: dict = pickle.load(fo, encoding='bytes') return dict qmnist = unpickle("MNIST-120k")

The data comes in a dictionary format, you can get the data and the labels separately by extracting the content from the dictionary: data = qmnist['data'] labels = qmnist['labels']

Source

The original QMNIST dataset was uploaded by Chhavi Yadav and Léon Bottou. Citation:

Yadav, C. and Bottou, L., “Cold Case: The Lost MNIST Digits”, arXiv e-prints, 2019.

Link to the original paper: https://arxiv.org/pdf/1905.10498.pdf Link to the GitHub repository: https://github.com/facebookresearch/qmnist

My contribution was to collect all the images and labels into the same file and convert it into a pickle file so it is easier to load. Please consider mentioning the author if you use this dataset instead of the original version.

Snapshot img

AI In Computer Vision Market Size 2025-2029

The ai in computer vision market size is valued to increase by USD 26.68 billion, at a CAGR of 17.8% from 2024 to 2029. Proliferation of advanced multimodal and generative AI models will drive the ai in computer vision market.

Market Insights

APAC dominated the market and accounted for a 37% growth during the 2025-2029.
By Component - Hardware segment was valued at USD 5.77 billion in 2023
By Application - Image identification segment accounted for the largest market revenue share in 2023

Market Size & Forecast

Market Opportunities: USD 388.59 million 
Market Future Opportunities 2024: USD 26675.50 million
CAGR from 2024 to 2029 : 17.8%

Market Summary

The computer vision market is experiencing a significant surge due to the proliferation of advanced multimodal and generative AI models. These models, capable of processing visual and textual data, are revolutionizing industries by enabling new applications and enhancing existing ones. One such application is supply chain optimization. By integrating computer vision with logistics operations, companies can automate the identification and sorting of goods, reducing manual labor and errors. However, this technological advancement also presents complex challenges. Navigating a complex and evolving regulatory and ethical landscape is crucial for businesses adopting computer vision technology. Ethical considerations, such as privacy concerns and potential biases, must be addressed to ensure transparency and fairness. Additionally, regulatory compliance is essential to mitigate risks and maintain trust with customers and stakeholders. Despite these challenges, the potential benefits of computer vision technology are substantial, making it a critical area of investment for businesses seeking operational efficiency and competitive advantage.

What will be the size of the AI In Computer Vision Market during the forecast period?

Get Key Insights on Market Forecast (PDF) Request Free SampleThe market continues to evolve, with advancements in technology driving innovation across various industries. One notable trend is the integration of multi-camera systems, enabling more comprehensive scene understanding and object tracking algorithms. According to recent research, the use of multiple cameras in computer vision applications has led to a 30% increase in accuracy compared to single-camera systems. This improvement can significantly impact business decisions, from compliance and safety in manufacturing to product strategy and customer experience in retail. Computational photography techniques, such as histogram equalization and image compression, play a crucial role in enhancing image quality and reducing bandwidth requirements. Neural network training and biometric authentication are essential components of AI-driven systems, ensuring sharpness enhancement, noise reduction, and activity recognition. Furthermore, cloud-based vision solutions and model compression methods facilitate edge computing deployment and real-time video analytics. Incorporating lossless and lossy image compression, synthetic data generation, and parallel processing frameworks, companies can optimize high-throughput image processing and gpu acceleration techniques. Model deployment strategies, such as image fusion algorithms and data annotation tools, streamline the development and implementation of AI systems. Visual tracking algorithms and scene understanding further enhance the capabilities of these systems, providing valuable insights for businesses.

Unpacking the AI In Computer Vision Market Landscape

In the dynamic realm of artificial intelligence (AI) in computer vision, advanced technologies such as object detection models and facial recognition systems continue to revolutionize industries. These solutions have demonstrated significant improvements in business outcomes, with object detection models achieving a 30% increase in production efficiency by automating quality control in manufacturing, and facial recognition systems reducing identity verification time by up to 50% in the financial sector.

Sophisticated techniques like satellite imagery processing, camera calibration, and depth image processing enable accurate analysis of vast datasets in various industries, including agriculture, infrastructure management, and urban planning. Furthermore, AI in computer vision encompasses cutting-edge applications such as visual question answering, medical image analysis, and autonomous vehicle perception.

To ensure optimal performance, computer vision APIs employ feature extraction methods, data augmentation strategies, and model training optimization techniques. Performance evaluation metrics, such as precision-recall curves, intersection over union, and F1 scores, provide valuable insights into model effectiveness.

The diverse appl

Snapshot img

Machine Vision (MV) Market Size 2024-2028

The machine vision (mv) market size is valued to increase USD 8.95 billion, at a CAGR of 10.87% from 2023 to 2028. Significant cost savings in operation due to process control will drive the machine vision (mv) market.

Major Market Trends & Insights

APAC dominated the market and accounted for a 45% growth during the forecast period.
By End-user - Industrial segment was valued at USD 7.45 billion in 2022
By Type - Vision system segment accounted for the largest market revenue share in 2022

Market Size & Forecast

Market Opportunities: USD 107.56 million
Market Future Opportunities: USD 8949.30 million
CAGR : 10.87%
APAC: Largest market in 2022

Market Summary

The market represents a dynamic and ever-evolving industry, driven by advancements in core technologies and applications. With significant cost savings in operation due to process control, machine vision systems are increasingly incorporating thermal inspection capabilities. Core technologies, such as artificial intelligence (AI) and machine learning (ML), continue to revolutionize the sector, enabling more accurate and efficient inspections. Applications span industries, from automotive to electronics, and service types range from hardware and software to consulting and integration. The market is intensely competitive, with key players vying for market share. For instance, according to a recent report, the global machine vision market is projected to reach a 21% share by 2027. Despite these opportunities, challenges such as high implementation costs and data security concerns persist. Regulatory compliance and regional differences also shape the market landscape.

What will be the Size of the Machine Vision (MV) Market during the forecast period?

Get Key Insights on Market Forecast (PDF) Request Free Sample

How is the Machine Vision (MV) Market Segmented and what are the key trends of market segmentation?

The machine vision (mv) industry research report provides comprehensive data (region-wise segment analysis), with forecasts and estimates in 'USD million' for the period 2024-2028, as well as historical data from 2018-2022 for the following segments. End-userIndustrialNon-industrialTypeVision systemCamerasOthersGeographyNorth AmericaUSCanadaEuropeGermanyAPACChinaJapanRest of World (ROW)

By End-user Insights

The industrial segment is estimated to witness significant growth during the forecast period.

Machine vision (MV) technology continues to revolutionize industries by enabling advanced visual feature extraction and analysis. Currently, the industrial sector dominates the market, with significant contributions from the automotive and electronics industries, as well as the metal and food and beverages sectors. The pharmaceutical industry is also increasing its adoption of machine vision systems, particularly for process line optimization and logistics applications. Looking forward, the market is poised for substantial growth, with the automotive industry projected to lead the charge. Additionally, the electronics sector, metal industry, and food and beverages industry are anticipated to witness substantial expansion. The manufacturing industry's competitive landscape is driving the demand for machine vision technology, as companies seek to enhance production rates and minimize product defects. Visual feature extraction, 3D vision processing, hyperspectral imaging analysis, multispectral imaging sensors, lidar data processing, autonomous vehicle navigation, defect detection algorithms, industrial automation systems, infrared imaging systems, thermal imaging applications, robotics vision guidance, pattern recognition software, facial recognition software, computer vision libraries, embedded vision systems, scene understanding methods, automated visual inspection, object detection systems, depth sensing technology, machine learning pipelines, camera calibration methods, real-time image processing, motion tracking algorithms, pose estimation techniques, optical character recognition, object classification accuracy, medical imaging diagnostics, image segmentation techniques, deep learning models, convolutional neural networks, image recognition algorithms, image enhancement filters, and gesture recognition technology are all integral components of the evolving machine vision market.

Request Free Sample

The Industrial segment was valued at USD 7.45 billion in 2018 and showed a gradual increase during the forecast period.

Request Free Sample

Regional Analysis

APAC is estimated to contribute 45% to the growth of the global market during the forecast period.Technavio’s analysts have elaborately explained the regional trends and drivers that shape the market during the forecast period.

See How Machine Vision (MV) Market Demand is Rising in APAC Request Fr

The file called analysis.pdf provides the code we used to analyze data. The filed called module.pdf is the module we created to access Numina data. The module requires the python packages called gql and pandas. The code requires a password from Numina.

The precise determination of leaf shape is crucial for the quantification of morphological variations between individual leaf ranks and cultivars and simulating their impact on light interception in functional-structural plant models (FSPMs). Standard manual measurements on destructively collected leaves are time-intensive and prone to errors, particularly in maize (Zea mays L.), which has large, undulating leaves that are difficult to flatten. To overcome these limitations, this study presents a new camera method developed as an image-based computer vision approach method for maize leaf shape analysis. A field experiment was conducted with seven commonly used silage maize cultivars at the experimental station Heidfeldhof, University of Hohenheim, Germany, in 2022. To determine the dimensions of fully developed leaves per rank and cultivar, three destructive measurements were conducted until flowering. The new camera method employs a GoPro Hero8 Black camera, integrated within an LI-3100C Area Meter, to capture high-resolution videos (1920 × 1080 pixels, 60 fps). A semi-automated software facilitates object detection, contour extraction, and leaf width determination, including calibration for accuracy. Validation was performed using pixel-counting and contrast analysis, comparing results against standard manual measurements to assess accuracy and reliability. Leaf width functions were fitted to quantify leaf shape parameters. Statistical analysis comparing cultivars and leaf ranks identified significant differences in leaf shape parameters (p < 0.01) for term alpha and term a. Simulations within a FSPM demonstrated that variations in leaf shape can alter light interception by up to 7%, emphasizing the need for precise parameterization in crop growth models. The new camera method provides a basis for future studies investigating rank-dependent leaf shape effects, which can offer an accurate representation of the canopy in FSPMs and improve agricultural decision-making.

This is a very large dataset from the CVPR 2019 publication, which can be found at https://arxiv.org/pdf/1812.01748.pdf. This contains a fashion-based dataset from Pinterest that offers the Shop the Look tool and suggests clothing based on a photograph of a certain situation. Each data set contains a pair of scenes and products.

Scenes: 47,739

Products: 38,111

Scene-Product Pairs: 93,274

The GitHub repository for this dataset can be found here.

NOTE: BY USING THIS DATASET YOU ACKNOWLEDGE THAT YOU HAVE READ THE LICENSE AND WILL ABIDE BY THE TERMS THEREWITHIN

THE LICENSE

ObjectNet is free to use for both research and commercial
applications. The authors own the source images and allow their use
under a license derived from Creative Commons Attribution 4.0 with
two additional clauses:

1. ObjectNet may never be used to tune the parameters of any
  model. This includes, but is not limited to, computing statistics
  on ObjectNet and including those statistics into a model,
  fine-tuning on ObjectNet, performing gradient updates on any
  parameters based on these images.

2. Any individual images from ObjectNet may only be posted to the web
  including their 1 pixel red border.

If you post this archive in a public location, please leave the password
intact as "objectnetisatestset".

[Other General License Information Conforms to Attribution 4.0 International]

This is Part 2 of 10 * Original Paper Link * ObjectNet Website

The links to the various parts of the dataset are:

• Part 1
• Part 2
• Part 3
• Part 4
• Part 5
• Part 6
• Part 7
• Part 8
• Part 9
• Part 10

Description From ObjectNET Homepage

What is ObjectNet?

• A new kind of vision dataset borrowing the idea of controls from other areas of science.
• No training set, only a test set! Put your vision system through its paces.
• Collected to intentionally show objects from new viewpoints on new backgrounds.
• 50,000 image test set, same as ImageNet, with controls for rotation, background, and viewpoint.
• 313 object classes with 113 overlapping ImageNet
• Large performance drop, what you can expect from vision systems in the real world!
• Robust to fine-tuning and a very difficult transfer learning problem

Controls For Biases Increase Variation

https://objectnet.dev/images/objectnet_controls_table.png">

Easy For Humans, Hard For Machines

• Ready to help develop the next generation of object recognition algorithms that have robustness, bias, and safety in mind.
• Controls can remove bias from other datasets machine learning, not just vision.

https://objectnet.dev/images/objectnet_results.png">

Full Description

ObjectNet is a large real-world test set for object recognition with control where object backgrounds, rotations, and imaging viewpoints are random.

Most scientific experiments have controls, confounds which are removed from the data, to ensure that subjects cannot perform a task by exploiting trivial correlations in the data. Historically, large machine learning and computer vision datasets have lacked such controls. This has resulted in models that must be fine-tuned for new datasets and perform better on datasets than in real-world applications. When tested on ObjectNet, object detectors show a 40-45% drop in performance, with respect to their performance on other benchmarks, due to the controls for biases. Controls make ObjectNet robust to fine-tuning showing only small performance increases.

We develop a highly automated platform that enables gathering datasets with controls by crowdsourcing image capturing and annotation. ObjectNet is the same size as the ImageNet test set (50,000 images), and by design does not come paired with a training set in order to encourage generalization. The dataset is both easier than ImageNet – objects are largely centred and unoccluded – and harder, due to the controls. Although we focus on object recognition here, data with controls can be gathered at scale using automated tools throughout machine learning to generate datasets that exercise models in new ways thus providing valuable feedback to researchers. This work opens up new avenues for research in generalizable, robust, and more human-like computer vision and in creating datasets where results are predictive of real-world performance.

Citation

...

The Moorea Labeled Corals dataset is a subset of the MCR LTER packaged for computer vision research. It contains 2055 images from three habitats IDs: fringing reef outer 10m and outer 17m, from 2008, 2009 and 2010. It also contains random point annotation (row, col, label) for the nine most abundant labels, four non coral labels: (1) Crustose Coralline Algae (CCA), (2) Turf algae, (3) Macroalgae and (4) Sand, and five coral genera: (5) Acropora, (6) Pavona, (7) Montipora, (8) Pocillopora, and (9) Porites. These nine classes account for 96% of the annotations and total to almost 400,000 points. These nine classes are the ones analyzed in (Beijbom, 2012); less-abundant genera not treated in the automation are also present in the dataset.

 These data were published in 
      Beijbom O., Edmunds P.J., Kline D.I., Mitchell G.B., Kriegman D., 'Automated Annotation of Coral Reef Survey Images', 
      IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Providence, Rhode Island, 2012.
      [BibTex]
      [pdf]


 These data are a subset of the raw data from which knb-lter-mcr.4 is derived.

Deep learning, a state-of-the-art machine learning approach, has shown outstanding performance over traditional machine learning in identifying intricate structures in complex high-dimensional data, especially in the domain of computer vision. The application of deep learning to early detection and automated classification of Alzheimer's disease (AD) has recently gained considerable attention, as rapid progress in neuroimaging techniques has generated large-scale multimodal neuroimaging data. A systematic review of publications using deep learning approaches and neuroimaging data for diagnostic classification of AD was performed. A PubMed and Google Scholar search was used to identify deep learning papers on AD published between January 2013 and July 2018. These papers were reviewed, evaluated, and classified by algorithm and neuroimaging type, and the findings were summarized. Of 16 studies meeting full inclusion criteria, 4 used a combination of deep learning and traditional machine learning approaches, and 12 used only deep learning approaches. The combination of traditional machine learning for classification and stacked auto-encoder (SAE) for feature selection produced accuracies of up to 98.8% for AD classification and 83.7% for prediction of conversion from mild cognitive impairment (MCI), a prodromal stage of AD, to AD. Deep learning approaches, such as convolutional neural network (CNN) or recurrent neural network (RNN), that use neuroimaging data without pre-processing for feature selection have yielded accuracies of up to 96.0% for AD classification and 84.2% for MCI conversion prediction. The best classification performance was obtained when multimodal neuroimaging and fluid biomarkers were combined. Deep learning approaches continue to improve in performance and appear to hold promise for diagnostic classification of AD using multimodal neuroimaging data. AD research that uses deep learning is still evolving, improving performance by incorporating additional hybrid data types, such as—omics data, increasing transparency with explainable approaches that add knowledge of specific disease-related features and mechanisms.

Humans and animals recognize objects irrespective of the beholder's point of view, which may drastically change their appearance. Artificial pattern recognizers strive to also achieve this, e.g., through translational invariance in convolutional neural networks (CNNs). However, CNNs and vision transformers (ViTs) both perform poorly on rotated inputs. Here we present AMR (artificial mental rotation), a method for dealing with in-plane rotations focusing on large datasets and architectural flexibility, our simple AMR implementation works with all common CNN and ViT architectures. We test it on randomly rotated versions of ImageNet, Stanford Cars, and Oxford Pet. With a top-1 error (averaged across datasets and architectures) of 0.743, AMR outperforms rotational data augmentation (average top-1 error of 0.626) by 19%. We also easily transfer a trained AMR module to a downstream task to improve the performance of a pre-trained semantic segmentation model on rotated CoCo from 32.7 to 55.2 IoU.

This package supplements the following paper submitted to ESSD: http://www.earth-syst-sci-data-discuss.net/essd-2017-7/essd-2017-7.pdf

This package contains all collected data that is available in a text format, as well as useful code. The following subfolders are containted within "floodX Datasets":

• metadata: Contains metadata for the various sensors, data source files, flooding experiments and image files for optical character recognition.
• data_raw: Contains the raw text format data, organized by sensor.
• code: Contains processing code for making the raw data more usable. The code is maintained in the following github repository: https://github.com/eawag-sww/floodx_data_preprocessing.git
• data_ocr_result: Contains data read from sensor display images using the processing code.
• data_preprocessed: Contains a version of the preprocessed code.

This repository contains the multiple instance learning datasets previously stored at miproblems.org. As I am now longer maintaining the website, I moved the datasets to Figshare. A detailed description of the files is found in readme.pdf

If you use these datasets, please cite this Figshare resource rather than linking to miproblems.org, which will be offline soon.

We present TransProteus, a dataset, for predicting the 3D structure and properties of materials, liquids, and objects inside transparent vessels from a single image without prior knowledge of the image source and camera parameters. Manipulating materials in transparent containers is essential in many fields and depends heavily on vision. This work supplies a new procedurally generated dataset consisting of 50k images of liquids and solid objects inside transparent containers. The image annotations include 3D models and material properties (color/transparency/roughness...) for the vessel and its content. The synthetic (CGI) part of the dataset was procedurally generated using 13k different objects, 500 different environments (HDRI), and 1450 material textures (PBR) combined with simulated liquids and procedurally generated vessels. In addition, we supply 104 real-world images of objects inside transparent vessels with depth maps of both the vessel and its content.

Note that there are two files here:

Transproteus_SimulatedLiquids2_New_No_Shift.7z

and

TranProteus2.7z , contain subset of the virtual CGI data set.

https://zenodo.org/api/files/12b013ca-36be-4156-afd4-c93b5fa22093/Tansproteus_SimulatedLiquids2_New_No_Shift.7z

TransProteus_RealSense_RealPhotos.7z : Contain real-world photos scanned with real sense with depth map of both the vessel and its content

See ReadMe file in side the downloaded files for more details

The full dataset (>100gb) can be found here:

https://e.pcloud.link/publink/show?code=kZfx55Zx1GOrl4aUwXDrifAHUPSt7QUAIfV

https://icedrive.net/1/6cZbP5dkNG

See: https://arxiv.org/pdf/2109.07577.pdf for more details

**This dataset is complementary to LabPics dataset with 8k real images of materials in vessels in chemistry labs, medical labs, and other settings. The LabPics dataset can be downloaded from here:

https://zenodo.org/record/4736111#.YVOAx3tE1H4

Transproteus_SimulatedLiquids2_New_No_Shift.7z and TranProteus2.7z

The two folders contain relatively similar data styles. The data in No_Shift contain images that were generated with no camera shift in the camera paramters. If you try to predict 3d model from an image as a depth map, this is easier to use (Otherwise, you need to adapt the image using the shift). For all other purposes, both folders are the same, and you can use either or both. In addition, a real image dataset for testing is given in the RealSense file.

The Robot-at-Home dataset (Robot@Home, paper here) is a collection of raw and processed data from five domestic settings compiled by a mobile robot equipped with 4 RGB-D cameras and a 2D laser scanner. Its main purpose is to serve as a testbed for semantic mapping algorithms through the categorization of objects and/or rooms.

This dataset is unique in three aspects:

• The provided data were captured with a rig of 4 RGB-D sensors with an overall field of view of 180°H. and 58°V., and with a 2D laser scanner.
• It comprises diverse and numerous data: sequences of RGB-D images and laser scans from the rooms of five apartments (87,000+ observations were collected), topological information about the connectivity of these rooms, and 3D reconstructions and 2D geometric maps of the visited rooms.
• The provided ground truth is dense, including per-point annotations of the categories of the objects and rooms appearing in the reconstructed scenarios, and per-pixel annotations of each RGB-D image within the recorded sequences

During the data collection, a total of 36 rooms were completely inspected, so the dataset is rich in contextual information of objects and rooms. This is a valuable feature, missing in most of the state-of-the-art datasets, which can be exploited by, for instance, semantic mapping systems that leverage relationships like pillows are usually on beds or ovens are not in bathrooms.

Robot@Home2

Robot@Home2, is an enhanced version aimed at improving usability and functionality for developing and testing mobile robotics and computer vision algorithms. It consists of three main components. Firstly, a relational database that states the contextual information and data links, compatible with Standard Query Language. Secondly,a Python package for managing the database, including downloading, querying, and interfacing functions. Finally, learning resources in the form of Jupyter notebooks, runnable locally or on the Google Colab platform, enabling users to explore the dataset without local installations. These freely available tools are expected to enhance the ease of exploiting the Robot@Home dataset and accelerate research in computer vision and robotics.

If you use Robot@Home2, please cite the following paper:

Gregorio Ambrosio-Cestero, Jose-Raul Ruiz-Sarmiento, Javier Gonzalez-Jimenez, The Robot@Home2 dataset: A new release with improved usability tools, in SoftwareX, Volume 23, 2023, 101490, ISSN 2352-7110, https://doi.org/10.1016/j.softx.2023.101490.

@article{ambrosio2023robotathome2,
title = {The Robot@Home2 dataset: A new release with improved usability tools},
author = {Gregorio Ambrosio-Cestero and Jose-Raul Ruiz-Sarmiento and Javier Gonzalez-Jimenez},
journal = {SoftwareX},
volume = {23},
pages = {101490},
year = {2023},
issn = {2352-7110},
doi = {https://doi.org/10.1016/j.softx.2023.101490},
url = {https://www.sciencedirect.com/science/article/pii/S2352711023001863},
keywords = {Dataset, Mobile robotics, Relational database, Python, Jupyter, Google Colab}
}

Version history
v1.0.1 Fixed minor bugs.
v1.0.2 Fixed some inconsistencies in some directory names. Fixes were necessary to automate the generation of the next version.
v2.0.0 SQL based dataset. Robot@Home v1.0.2 has been packed into a sqlite database along with RGB-D and scene files which have been assembled into a hierarchical structured directory free of redundancies. Path tables are also provided to reference files in both v1.0.2 and v2.0.0 directory hierarchies. This version has been automatically generated from version 1.0.2 through the toolbox.
v2.0.1 A forgotten foreign key pair have been added.
v.2.0.2 The views have been consolidated as tables which allows a considerable improvement in access time.
v.2.0.3 The previous version does not include the database. In this version the database has been uploaded.
v.2.1.0 Depth images have been updated to 16-bit. Additionally, both the RGB images and the depth images are oriented in the original camera format, i.e. landscape.

The ongoing need to sustainably manage fishery resources can benefit from fishery-independent monitoring of fish stocks. Camera systems, particularly baited remote underwater video system (BRUVS), are a widely used and repeatable method for monitoring relative abundance, required for building stock assessment models. The potential for BRUVS-based monitoring is restricted, however, by the substantial costs of manual data extraction from videos. Computer vision, in particular deep learning (DL) models, are increasingly being used to automatically detect and count fish at low abundances in videos. One of the advantages of BRUVS is that bait attractants help to reliably detect species in relatively short deployments (e.g., 1 h). The high abundances of fish attracted to BRUVS, however, make computer vision more difficult, because fish often obscure other fish. We build upon existing DL methods for identifying and counting a target fisheries species across a wide range of fish abundances. Using BRUVS imagery targeting a recovering fishery species, Australasian snapper (Chrysophrys auratus), we tested combinations of three further mathematical steps likely to generate accurate, efficient automation: (1) varying confidence thresholds (CTs), (2) on/off use of sequential non-maximum suppression (Seq-NMS), and (3) statistical correction equations. Output from the DL model was more accurate at low abundances of snapper than at higher abundances (>15 fish per frame) where the model over-predicted counts by as much as 50%. The procedure providing the most accurate counts across all fish abundances, with counts either correct or within 1–2 of manual counts (R2 = 88%), used Seq-NMS, a 45% CT, and a cubic polynomial corrective equation. The optimised modelling provides an automated procedure offering an effective and efficient method for accurately identifying and counting snapper in the BRUV footage on which it was tested. Additional evaluation will be required to test and refine the procedure so that automated counts of snapper are accurate in the survey region over time, and to determine the applicability to other regions within the distributional range of this species. For monitoring stocks of fishery species more generally, the specific equations will differ but the procedure demonstrated here could help to increase the usefulness of BRUVS.

Global Blockchain AI Market size was valued at USD 348.57 Million in 2022 and is poised to grow from USD 430.83 Million in 2023 to USD 2346.68 Million by 2031, growing at a CAGR of 23.6% in the forecast period (2024-2031).

Abstract

Inspired by the recent success of RGB-D cameras, we propose the enrichment of RGB data with an additional quasi-free modality, namely, the wireless signal emitted by individuals' cell phones, referred to as RGB-W. The received signal strength acts as a rough proxy for depth and a reliable cue on a person's identity. Although the measured signals are noisy, we demonstrate that the combination of visual and wireless data significantly improves the localization accuracy. We introduce a novel image-driven representation of wireless data which embeds all received signals onto a single image. We then evaluate the ability of this additional data to (i) locate persons within a sparsity-driven framework and to (ii) track individuals with a new confidence measure on the data association problem. Our solution outperforms existing localization methods. It can be applied to the millions of currently installed RGB cameras to better analyze human behavior and offer the next generation of high-accuracy location-based services.

Conference Paper

PDF: http://www.cv-foundation.org/openaccess/content_iccv_2015/papers/Alahi_RGB-W_When_Vision_ICCV_2015_paper.pdf

Metadata

+----------------+-----------------+-----------+-----------+--------------+----------+ | Sequence Name | Length (mm:ss) | # Frames | # People | # W Devices | Download | +----------------+-----------------+-----------+-----------+--------------+----------+ | conference-1 | 01:53 | 1,697 | 5 | 5 | 116 MiB | | conference-2 | 05:18 | 4,782 | 12 | 12 | 379 MiB | | conference-3 | 23:31 | 21,165 | 1 | 2 | 1.3 GiB | | conference-4 | 06:27 | 4,832 | 1 | 2 | 357 MiB | | conference-5 | 06:03 | 4,525 | 2 | 2 | 290 MiB | | patio-1 | 07:22 | 6,636 | 4 | 4 | 474 MiB | | patio-2 | 04:36 | 4,144 | 2 | 2 | 258 MiB | | Full Dataset | 55:10 | 47,781 | -- | -- | 3.2 GiB | +----------------+-----------------+-----------+-----------+--------------+----------+

Citation

If you would like to cite our work, please use the following.

Alahi A, Haque A, Fei-Fei L. (2015). RGB-W: When Vision Meets Wireless. International Conference on Computer Vision (ICCV). Santiago, Chile. IEEE.

@inproceedings{alahi2015rgb, title={RGB-W: When vision meets wireless}, author={Alahi, Alexandre and Haque, Albert and Fei-Fei, Li}, booktitle={International Conference on Computer Vision}, year={2015} }

The "Pelerinage" dataset contains a fine-grained edition of excerpts from 20 medieval manuscripts of Guillaume de Digulleville's Pelerinage de vie humaine.

Files in this directory were created as part of the ECMEN Ecritures médiévales et outils numériques research project, funded by the City of Paris in the Emergence(s) framework.

If you use any of the following files, please quote:

Stutzmann, Dominique. « Les 'manuscrits datés', base de données sur l’écriture ». In Catalogazione, storia della scrittura, storia del libro. I Manoscritti datati d’Italia vent’anni dopo, ed. Teresa De Robertis and Nicoletta Giovè Marchioli, Firenze: SISMEL - Edizioni del Galluzzo, 2017, p. 155-207.

@incollection{stutzmann_les_2017,
  address = {Firenze},
  title = {Les « manuscrits datés », base de données sur l’écriture},
  language = {fre},
  booktitle = {Catalogazione, storia della scrittura, storia del libro. {I} {Manoscritti} datati d’{Italia} vent’anni dopo},
  publisher = {SISMEL - Edizioni del Galluzzo},
  author = {Stutzmann, Dominique},
  editor = {De Robertis, Teresa and Giovè Marchioli, Nicoletta},
  year = {2017},
  pages = {155--207}
}

SOURCE DESCRIPTION

The source of the files are imitative transcriptions of selected passages by Géraldine Veysseyre (ORCID 0000-0002-3737-2137) based on 20 medieval manuscripts containing the "Pelerinage de vie humaine" of Guillaume de Digulleville, as produced during the OPVS research project (the acronym stands for « Old Pious Vernacular Successes » in English and « Œuvres Pieuses Vernaculaires à Succès » in French), funded by the ERC under the grant agreement n° 263274).

The transcriptions were edited and enhanced in TEI format by Dominique Stutzmann (ORCID 0000-0003-3705-5825) and Floriana Ceresato (IRHT-CNRS), as part of the ORIFLAMMS and the ECMEN (Ecritures médiévales et outils numériques) research project with following features:

• semantic encoding (