Executive Summary : | Baked goods, such as bread and sponge cakes, have been a staple food for centuries due to their quality, taste, and visual appeal. However, early spoilage detection has not fully realized its potential. Traditional methods, such as molecular-level analysis (MLA) techniques, have been limited in their detection. Hyperspectral imaging (HsI) has been explored for detecting spoilage in food due to mold, bacteria, and other microbes. However, there is a lack of HsI research on baked foods. This study presents a unique filter-free multispectral reflectance photography-based HsI approach for automatically detecting and predicting microbiological deterioration in bread items. A spectral preserve fusion technique will be developed to enhance the spatial quality of HsI photographs without altering the spectral data. Artificial intelligence (AI)-based methods, including deep learning and support vector machine, will be applied to automate the detection process. The proposed technology is expected to detect spoilage at least 24 hours before it is visible to the naked eye. The trained model's efficacy will be checked using a separate dataset. |
Co-PI: | Dr. sivaraman J, National Institute Of Technology (NIT) Rourkela, Odisha-769008, Dr. Kasturi Dutta, National Institute Of Technology (NIT) Rourkela, Odisha-769008, Dr. Bala Chakravarthy Neelapu, National Institute Of Technology (NIT) Rourkela, Odisha-769008 |