Abstract:
Deep learning replicate human brain to help the system to solve complex problem like
identifying object. In this researched based project, we used deep learning to identify cow
species. To know something, human often depend on the technology like object
classification. People of today’s generation use software like google lens to identify the
unknown. Among all domestic animals, Cow is a the most common and useful around us.
Based on their species, they are useful to different need. Cow provides meat, milk and etc.
So, by this research we used deep learning on a data set for identification of cow species.
The date set is created by collecting photos using mobile phone and as accurate as possible.
There is total of seven species and around two thousands of raw data. We used python to
resized the data set in zip file as a part of data preprocessing. I used ResNet50, ResNet152,
DenseNet121, and DenseNet201 and compare them to get the most accuracy. Among them
DenseNet201 perform max and gave the accuracy of 97.35. According to my background
study, this result is maximum on cow spices of local area. My research will inspire the new
coming researcher to work agriculture sector. The research finding will contribute in the
future digital firming. Using a collection of photos, this study investigated the effectiveness
of transfer learning approaches for the classification of seven different cow species. The
goal of the study was to classify cow species with the best possible accuracy.
