Research

Computer Vision

Announcing the winners of the Computer Vision for Global Challenges research awards

October 11, 2019

Recent advancements in the field of computer vision (CV) have led to new applications that could benefit people globally, and especially those in developing countries. To bring the CV community closer to tasks, datasets, and applications that can have a global impact, Facebook AI launched the Computer Vision for Global Challenges (CV4GC) initiative earlier this year. Through a series of academic programs, mentorships, sponsorships, and events, CV4GC brings together field experts from around the world to discuss potential CV applications to address issues that affect developing regions.

One such program is the CV4GC request for proposals, a research award opportunity that launched in February with the goal of supporting research that aligns with CV4GC’s mission. We were particularly interested in proposals that extended CV technology to achieve global development priorities, especially those captured in the United Nations’ Sustainable Development Goals. Three winning proposals have been chosen and are listed below, along with brief research summaries. (Proposal summaries have been slightly adapted from submissions).

We received more than 300 applications, each of which was reviewed twice by the judging committee. “We’re committed to making sure AI works for everyone, no matter who they are or where they live,” says Andrew Westbury, Research Program Manager for the CV4GC initiative. “The winning proposals from academics in India, Pakistan, and Colombia will help the industry broaden how they approach CV and possibly lead to world-changing benefits for society.”

Thank you to everyone who took the time to submit a proposal, and congratulations to the winners.

Research award winners

AnnaData: An early warning system for crop diseases

PI: Santosh Kesavan, Crosslinks Foundation

As per FAO (UN Food and Agriculture Organization) estimates, plant pests and diseases account for approximately 30 percent of global crop production losses across the world. Since plant-based food forms a major portion of the human diet worldwide, plant pests and diseases pose a major threat to global food security and crop losses by severely reducing the availability and access to plant food. The objective of this proposal is to design, develop, and commercialize a user-friendly and low-cost smartphone-based technology and sensor system called AnnaData for Indian farmers that does real-time processing of comprehensive weather, plant, and field data gathered through on-site sensors.

Low-cost deep learning solution to real-time detection of malaria

PI: Waqas Sultani, Information Technology University

Malaria is a life-threatening disease that is caused by infected female Anopheles mosquitoes. According to the World Health Organization, 219 million cases of malaria were observed in 87 countries in 2017 alone, and 435,000 people lost their lives due to the disease. Malaria is curable, and early detection can save lives. Malaria, also named as a disease of the poor, mostly happens in low-income rural areas. For them, getting a laboratory test (using an expensive microscope) and getting it checked by a trained medical doctor is very difficult. This research proposal aims to provide a low-cost solution for early detection of malaria patients through the detection of the malaria parasite in microscopic blood-sample images.

Portable device to analyze thick blood smears for malaria diagnosis

PI: Wendy Fong, Pontificia Universidad Javeriana

A timely diagnosis and immediate effective treatment are the bases for the management of malaria to reduce the morbidity and mortality caused by this disease. Malaria is one of the most serious public health problems in the world. This project proposes a system for improving diagnosis time and diagnosis quality. The proposed system is a portable device that can adapt to the microscopes used in the field and help microscopists not only analyze the quality of the smear, but also aid in the diagnosis of malaria by automatic calculation of parasitic density. With the system, it will be possible to visually follow patient history and evolution to treatment.

To view research awards currently open for submissions and to subscribe to our email list, visit our Research Awards page.