The State of Computer Vision Research in Africa

TL;DR

This paper analyzes the state of computer vision research in Africa by combining a bottom-up catalog of African CV datasets with topic and collaboration analyses and a large-scale researcher questionnaire. Using three data scopes (full, refined, top-tier) and LLM-assisted classification, it documents 96 officially published and 33 unofficial datasets across 31 categories, and identifies dominant regional topics and inequities in global publishing access. Key findings include Africa's 0.06% share of top-tier CV publications, pronounced regional disparities, and a strong call for intra-African collaborations and curating local datasets. The work advances a decolonial, participatory framework for CV in Africa and provides actionable resources and directions for capacity-building, policy, and training initiatives.

Abstract

Despite significant efforts to democratize artificial intelligence (AI), computer vision which is a sub-field of AI, still lags in Africa. A significant factor to this, is the limited access to computing resources, datasets, and collaborations. As a result, Africa's contribution to top-tier publications in this field has only been 0.06% over the past decade. Towards improving the computer vision field and making it more accessible and inclusive, this study analyzes 63,000 Scopus-indexed computer vision publications from Africa. We utilize large language models to automatically parse their abstracts, to identify and categorize topics and datasets. This resulted in listing more than 100 African datasets. Our objective is to provide a comprehensive taxonomy of dataset categories to facilitate better understanding and utilization of these resources. We also analyze collaboration trends of researchers within and outside the continent. Additionally, we conduct a large-scale questionnaire among African computer vision researchers to identify the structural barriers they believe require urgent attention. In conclusion, our study offers a comprehensive overview of the current state of computer vision research in Africa, to empower marginalized communities to participate in the design and development of computer vision systems.

Figures (10)

• Figure 1: Our proposed pipeline for data collection, verification, and analysis of the African Scopus-indexed computer vision publications. The search query generation uses simple queries to retrieve all computer vision publications (i.e., full set) or generates queries based on the top-50 keywords in computer vision as a sample (i.e., refined set). This is followed by data collection of the full, refined and top-tier publications sets and a verification phase on the refined and top-tier sets. Finally, we perform classification and analysis combining automatic, i.e., large language models parsing abstracts, and manual categorization.
• Figure 2: Taxonomy of the datasets categories for the retrieved officially published datasets and unofficial ones hosted in challenges and data host platforms. We show, "category: number of datasets retrieved under this category".
• Figure 3: Topic categories and keywords taxonomies of the retrieved publications. The first level is the topic category, while the second level shows the keywords that are categorized under that topic.
• Figure 4: Keywords analysis of the top-30 recurring keywords. (A-E) Distribution of the most recurring keywords per African region. Red indicates the percentage of publications within the corresponding region indexed with that keyword. We remove four words from the top-30 set corresponding to general topics in computer vision (i.e., Computer Vision, Camera, Deep Learning, Convolutional Neural Networks) to focus on fine-grained topics.
• Figure 5: Scopus-indexed computer vision publications per African region across the time interval 2012-2022 showing the number of publications. We use the logarithmic scale. It shows consistent growth in Northern and Southern regions’ publications, and a recent increase in Eastern and Western Africa (2016-2022). However, Central Africa is the most in need of improving the computer vision capacity.