Computer Vision Services Research Articles

Shifting Our Awareness, Taking Back Tags: Temporal Changes in Computer Vision Services’ Social Behaviors

Much attention has been on the behaviors of computer vision services when describing images of people. Audits have revealed rampant biases that could lead to harm, when services are used by developers and researchers. We focus on temporal auditing, replicating experiments originally conducted three years ago. We document the changes observed over time, relating this to the growing awareness of structural oppression and the need to align technology with social values. While we document some positive changes in the services’ behaviors, such as increased accuracy in the use of gender-related tags overall, we also replicate findings concerning larger error rates for images of Black individuals. In addition, we find cases of increased use of inferential tags (e.g., emotions), which are often sensitive. The analysis underscores the difficulty in following changes in services’ behaviors over time, and the need for more oversight of such services.

Software for generation of video files metadata

In today’s world, information is one of the most important resources. Its large turnover in the media makes companies look for ways to automate processes related to labeling and storage of data. But often these processes are left on the shoulders of archive workers who have to review the footage in real-time and create metadata based on what they see, which will be transferred into the archive. This process is often very time-consuming. As a result of the analysis of approaches to video data processing, comparative analysis of the efficiency and availability of computer vision services, a software for searching faces, and forming metadata in the video archive of the VGTRK company was designed and implemented.

A Survey on “Object Detection Algorithms for Visually Impaired People”

Vision is one in every of the foremost necessary human senses and it plays a really necessary role in human interaction with the surrounding objects. Until now many papers have been published on these topics that shows various different computer vision products and services by developing new electronic devices for the visually disabled people. The aim is to study different object detection methods. As compared to other Object detection methods, YOLO method has multiple advantages. In alternative algorithms like CNN, Fast-CNN the algorithmic program won't investigate the image fully however in YOLO the algorithmic program investigate the image fully by predicting the bounding boxes by making use of convolutional network and possibilities for these boxes and detects the image quicker as compared to alternative algorithms.

Requirements of API Documentation: A Case Study into Computer Vision Services

Using cloud-based computer vision services is gaining traction, where developers access AI-powered components through familiar RESTful APIs, not needing to orchestrate large training and inference infrastructures or curate/label training datasets. However, while these APIs seem familiar to use, their non-deterministic run-time behaviour and evolution is not adequately communicated to developers. Therefore, improving these services' API documentation is paramount-more extensive documentation facilitates the development process of intelligent software. In a prior study, we extracted 34 API documentation artefacts from 21 seminal works, devising a taxonomy of five key requirements to produce quality API documentation. We extend this study in two ways. Firstly, by surveying 104 developers of varying experience to understand what API documentation artefacts are of most value to practitioners. Secondly, identifying which of these highly-valued artefacts are or are not well-documented through a case study in the emerging computer vision service domain. We identify: (i) several gaps in the software engineering literature, where aspects of API documentation understanding is/is not extensively investigated; and (ii) where industry vendors (in contrast) document artefacts to better serve their end-developers. We provide a set of recommendations to enhance intelligent software documentation for both vendors and the wider research community.

Automatized analysis of children's exposure to child-directed speech in reschool settings: Validation and application.

The present study explored whether a tool for automatic detection and recognition of interactions and child-directed speech (CDS) in preschool classrooms could be developed, validated, and applied to non-coded video recordings representing children's classroom experiences. Using first-person video recordings collected by 13 preschool children during a morning in their classrooms, we extracted high-level audiovisual features from recordings using automatic speech recognition and computer vision services from a cloud computing provider. Using manual coding for interactions and transcriptions of CDS as reference, we trained and tested supervised classifiers and linear mappings to measure five variables of interest. We show that the supervised classifiers trained with speech activity, proximity, and high-level facial features achieve adequate accuracy in detecting interactions. Furthermore, in combination with an automatic speech recognition service, the supervised classifier achieved error rates for CDS measures that are in line with other open-source automatic decoding tools in early childhood settings. Finally, we demonstrate our tool's applicability by using it to automatically code and transcribe children's interactions and CDS exposure vertically within a classroom day (morning to afternoon) and horizontally over time (fall to winter). Developing and scaling tools for automatized capture of children's interactions with others in the preschool classroom, as well as exposure to CDS, may revolutionize scientific efforts to identify precise mechanisms that foster young children's language development.

COMPARATIVE ANALYSIS OF CLOUD COGNITIVE SERVICE PROVIDERS BY USING COMPUTER VISION’S OBJECT DETECTION

The given paper’s “Comparative analysis of cloud cognitive service providers by using computer vision’s object detection” goal is to analyze prediction precision of objects in digital image, using different cognitive services what are available in the market. During the research process there were made a several trials using different computer vision services to analyze the same number of images. Every separate image was under detailed analysis for prediction accuracy of each object. The results show that there is significant difference between different vision services.

Research and application of artificial intelligence service platform for the power field

Conventional analysis methods cannot fully meet the business needs of power grids. At present, several artificial intelligence (AI) projects in a single business field are competing with each other, and the interfaces between the systems lack unified specifications. Therefore, it is imperative to establish a comprehensive service platform. In this paper, an AI platform framework for power fields is proposed; it adopts the deep learning technology to support natural language processing and computer vision services. On one hand, it can provide an algorithm, a model, and service support for power- enterprise applications, and on the other hand, it can provide a large number of heterogeneous data processing, algorithm libraries, intelligent services, model managements, typical application scenarios, and other services for different levels of business personnel. The establishment of the platform framework could break data barrier, improve portability of technology, avoid the investment waste caused by repeated constructions, and lay the foundation for the construction of “platform + application + service” ecological chain.

Московский эксперимент по применению компьютерного зрения в лучевой диагностике: вовлеченность врачей-рентгенологов

B a c k g r o u n d . In 2019, the Moscow Government decided to conduct a large-scale scientific research – an the Experiment on the use of innovative computer vision technologies for medical image analysis and subsequent applicability in the healthcare system of Moscow (www.mosmed.ai). O b j e c t i v e – analyze engagement, attitudes and feedback from doctors-radiologists in frame of the Experiment. M a t e r i a l s a n d m e t h o d s . The Experiment is a prospective research approved by the Independent Ethics Committee and registered with Clinicaltrails.gov (ID NCT04489992). Patients signed informed voluntary consent. On the date 01.10.2020, ten services are involved in the Experiment, they providing automated analysis of chest computed tomography and x-ray, mammography. The study includes quantitative indicators of the Experiment from 06/18/2020 to 10/01/2020. Methods of social survey, descriptive statistics, assessment of diagnostic accuracy metrics were used. R e s u l t s a n d d i s c u s s i o n . During the first four months of the active phase of the Experiment, ten computer vision services were integrate into Unified Radiology Service of Moscow. More then 497 thousand studies have been successfully analyzed. Analyzes is carried out for 884 diagnostic devices in 293 medical organizations, 272 of them are actively involved. The involvement of medical organizations is 82%. The median time for automatic analysis of 1 study is 8 minutes. Overall, 63% of studies were analyzed in less than 15 minutes. At the beginning of the Experiment, 538 doctors had access to the system; in four months this number increased to 899. The involvement of doctors was 24%, which is slightly higher than the global indicators. According to the results of a sociological survey, the attitude to AI technologies of Moscow radiologists can be characterize as expectant, moderately optimistic. Radiologists have determined that the results of computer vision services are fully consistent with the real situation in 64% of cases. In 36% cases some inconsistencies were recorded; of this number, significant discrepancies took place in 6%, insignificant – in 23%. C o n c l u s i o n . Results of the Experiment’s first four months can be consider as successful. A high level of involvement of radiologists is define. Special measures will be implement to increase the involvement of radiologists, as well as a comprehensive comparative assessment of the work of services at the further stages of the Experiment.

How Computers See Gender

Investigations of facial analysis (FA) technologies-such as facial detection and facial recognition-have been central to discussions about Artificial Intelligence's (AI) impact on human beings. Research on automatic gender recognition, the classification of gender by FA technologies, has raised potential concerns around issues of racial and gender bias. In this study, we augment past work with empirical data by conducting a systematic analysis of how gender classification and gender labeling in computer vision services operate when faced with gender diversity. We sought to understand how gender is concretely conceptualized and encoded into commercial facial analysis and image labeling technologies available today. We then conducted a two-phase study: (1) a system analysis of ten commercial FA and image labeling services and (2) an evaluation of five services using a custom dataset of diverse genders using self-labeled Instagram images. Our analysis highlights how gender is codified into both classifiers and data standards. We found that FA services performed consistently worse on transgender individuals and were universally unable to classify non-binary genders. In contrast, image labeling often presented multiple gendered concepts. We also found that user perceptions about gender performance and identity contradict the way gender performance is encoded into the computer vision infrastructure. We discuss our findings from three perspectives of gender identity (self-identity, gender performativity, and demographic identity) and how these perspectives interact across three layers: the classification infrastructure, the third-party applications that make use of that infrastructure, and the individuals who interact with that software. We employ Bowker and Star's concepts of "torque" and "residuality" to further discuss the social implications of gender classification. We conclude by outlining opportunities for creating more inclusive classification infrastructures and datasets, as well as with implications for policy.