Fecal-based Health Status Prediction in Poultry Birds: A Blended Approach using Handcrafted and Deep Features

The well-being of animals and the profitability of poultry farming are both directly affected by the condition of the birds raised for food. Assessing the health of chickens has traditionally included a lot of physical effort and time-consuming procedures. In this study, a novel method is proposed for determining the health of chickens using fecal samples. Extraction of deep features from Convolutional Neural Networks (CNNs) is combined with the strengths of handcrafted features such as textual feature extraction and the Bag-of-Visual-Word (BoVW) proposed technique. The proposed hybrid model is meant to leverage the complementing traits of handmade and deep features to boost the accuracy of health status prediction. The proposed approach intends to provide a strong and efficient answer for poultry health evaluation by integrating information retrieved from the texture and visual content of fecal samples. Extensive tests were run, and hybrid model was compared to well-known deep architectures like Inception, ResNet50, and EfficientNet, to analyze the efficiency of the proposed model. The findings show that the proposed approach is superior to the others, with a remarkable 96 percent accuracy. This outstanding performance demonstrates the promise of proposed blended approach to greatly enhance the precision and efficiency of health status prediction in poultry birds, making it a useful resource for the poultry industry and animal health specialists.