RIASSUNTO
We propose an application paradigm of mobile cloud computing with big video data for ornamental fish warehousing (OFWare) system dedicated to koi fishes. In fields of aquaculture and agriculture, live creatures are the main products and are difficult to perform warehouse management. Warehousing of high unit-price ornamental fishes such as koi, stingray, and arowana is even more difficult since the warehousing requires identification, in addition to simple counting and classification, of individuals whose shapes and texture patterns are time-variant as they grow. Therefore, rather than using invasive RFID-based systems, we combine mobile cloud computing and big data analytics techniques including image and video collecting and transmission by handheld mobile devices, unsupervised texture pattern classification in fish tank videos, image retrieval based on support vector machine, and statistical analysis for the warehouse management. The proposed system is scalable based on Hadoop framework and a small unit of a single name-node and data-node is able to identify a fish feature among 500,000 koi fishes in 7 seconds.