Systematic Literature Review of Deep Learning models for Computer Vision Applications: Deployment Challenges in Nigeria

Abstract

Deep learning has gained attention recently. Since its adoption, deep learning has provided state-of-the-art solutions to lots of standing computational problems. One of the areas it has gained unequaled success is computer vision. The success of deep learning is not limited to computer vision only, it has also recorded unmatchable success in areas like natural language processing and speech recognition. With the advent of big data, the use and importance of deep learning can only continue to grow. One downside of this algorithm is its computational requirements: Literature Review, Developing countries large datasets and high-end computing devices. In this paper, we provide an overview of recent deep learning models for computer vision, and we also highlighted the challenges faced by developing countries in adopting these technologies. No review has covered the challenges faced by Nigeria in deploying this technology. Some of the challenges highlighted include manual data collection and lack of adequate cloud storage services. Inadequate infrastructures such as power and network facilities, and finally, lack of adequate funding of the sector. It was recommended that local cloud services be established to encourage local data storage and reduce storage cost. Also, adequate investment for power and network availability should be made. Finally, there should be enough budget allocation to IT sector that will encourage technocrat and experts to develop and fully harness the benefit of the technology