“Contrary to popular belief that all cattle look alike, this paper presents a current state of the art research and study in animal biometric based recognition a system which provides an important insight in the identification of cattle based on their facial images.”
– so explain researchers Santosh Kumar, Shrikant Tiwari, and Sanjay Kumar Singh of the Department of Computer Engineering, Indian Institute of Technology (BHU), and the Department of Computer Science and Engineering, Faculty of Engineering and Technology, Shri Shankaracharya Group of Institution, Shri Shankaracharya Technical Campus, Junwani, District-Durg, Bhilai, Chattisgarh, India, in a new paper for the journal Proceedings of the National Academy of Sciences, India Section A: Physical Sciences, June 2016, Volume 86, Issue 2, pp 137–148. It’s entitled ‘Face Recognition of Cattle: Can it be Done?’
The team created an image database of 3000 images (300 subjects each with 10 photos) and tested several different computerised methodologies, leading to an answer for their question which, broadly speaking, is ‘Yes’.
“The appearance (holistic) based face recognition approaches, independent component analysis (ICA) [36–38] algorithm yield the recognition accuracy of 86.95 % at the starting level of Gaussian smoothing. The PCA-LiBSVM [17–19, 32, 33, 41, 43] and ICA-LiBSVM [36–38, 41, 43] face recognition approaches provide the recognition accuracy of 95.62 and 95.87 % respectively.”
COMING SOON Other methods of computerized cow recognition that can also be done.
BONUS (facial recog. related): The privacy goggles of Prof. Echizen