Analysis of DenseNet -MobileNet-CNN Models on Image Classification using Bird Species Data
Manukonda Yaswanth Reddy, Mohammed Sajid Ahmed, Abhisyanth Nagumothu, Modepalli Kavitha · 2023
Image classification using a convolution neural network (CNN) has been one of the emerging topics in recent accomplishments. A good number of classification models are designed based on the CNN mechanism. Out of which, three deep neural network models are discussed in this article on image classification using bird species data. Bird species dataset is collected from public repository and deep neural network models are applied on it to classify the type of bird as part of implementation. The analysis of outcome produced by the model’s evidence that the Mobile Net model is given better result compared to Dense Net and traditional CNN models in terms of accuracy. This project will be looking at various entities which we discussed and will be likely experimenting with how good the model is in terms of recognition. There are many species that are endangered and there should be taken special care for their sustainment in the ecosystem. More than 10,000 species of these birds differ in their color, size, breed, and habitat part of the ecosystem. Mainly, this model focuses on image processing of the bird which is provided as an input to the model. CNN, Mobil Net, and Dense Net are such models that are very much challenging research areas where if there is even a slight change in a particular image, it will probably result in an entirely new image that a completely different species of bird might get generated. So, we will be thoroughly comparing the above three neural networks and provide the right amount of accuracy.