Published Paper Details:

Birds or Aves are a group of animals having a developed backbone which evolved from dinosaur. The Birds would be more precisely and scientifically defined as a group of warm-blooded animals having vertebrate column more related to reptile than mammals and they are possessing four chambered heart with forelimbs modified to wings, with lower developed sense of smell and keen vision with limited auditory range and they mostly reproduce by sexual means and by laying eggs. At present there are about 10,400 living species of bird each one is unique in having feathers, which is the major characteristic that distinguish them from all other animals. More than 1,000 bird species have been identified to be extinct and the number is increasing rapidly. Since earliest times, birds have been not only a animal category or a material but also a cultural resource where most have been divided and further re-divided by various means depending upon various features such as color, wings, modifications and habitat. At present, there are various species of bird which are at their edge of extinction and as of the newer generation getting low natural contact it become hard to identify a species for them. So, this study has come up with idea of a real time method called Bird Species Identification System. The method uses machine learning technique to create a model based on transfer learning system and implementing it as a web base application using Stream-lit. The method provides nearly 95 % accuracy for the 400 classes of birds. Thus, this study will favor machine learning based image classification unit with fair user interface which greatly involves in development of teaching, learning and visualizing process with Bird species recognition and obviously, the system will be capable to resolve the future challenges that may arise in that regards with Bird species identification and classification.