Prediction of Breed Lineage for Small Ruminant Production using Deep Learning Technique

According to UN Committee on World Food Security, people must always have access to sufficient food supplies such as meat, chicken, and sheep. Although important to Malaysian Muslims which account for about 60% of the population, sheep are in short supply locally due to the high mortality rate caused by fatal diseases such as Foot and Mouth Disease (FMD), Tetanus, etc. Because of inbreeding, qualities such as disease resistance, fertility, prolificacy, vigor, and survivability are reduced in animals, often referred to as inbreeding depression. It is important to note that infected sheep may cause contaminated sheep meat produce, transmitting foodborne bacteria such as E.coli and Salmonella to humans during different stages of food preparation. Previously, other papers compared hundreds of images of sheep to deep learning models to learn of its breed. Although successful, both methods took a long period to complete. This paper proposes a framework based on deep learning techniques that will identify and predict breed lineage and inherited disease in sheep. The adopted deep learning algorithm will improve time efficiency in retrieving immediate information while still maintaining a high accuracy rate. From a wider perspective, the proposed framework has the potential to be used across domains as it can be trained with any other dataset.