• Shwetha A
• Reshma M

In this project we use Deep Learning methods well-known as deep structured learning or deep machine learning based methods, this technique is based on artificial neural network that continuously receives the data and increases the efficiency during the training process for the detection and classification of fish using under water images. In this proposed method we have trained around 1500 images to classify by using Fast R-CNN method by this method we can obtain around 94% of accuracy. Deep learning methods have achieved excellent progress in digital image processing for fish detection and classification by means of Fast R-CNN (Regional and Convolution Neural Networks). This progress has motivated the requirement for automatic detection and classification of fish and also it will classify the types of fish within the fishes in the image data by using deep neural network based classifiers. This approach highlights the object of detection, deep learning architecture and the feature extraction methods used. And also we have extracted the features of fish and calculate the area and weights of the fish. The whole processes have been automated and it is resolved that there is great opportunity for automation in the identification of fish and classifying it, using deep neural network.

Deep learning; Machine learning; Fast R-CNN; Detecting and classifying.

"Deep Learning on Underwater Marine Fish Detection and Classification", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.5, Issue 8, page no.547-552, August-2018, Available :http://www.jetir.org/papers/JETIR1808557.pdf

"Deep Learning on Underwater Marine Fish Detection and Classification", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.5, Issue 8, page no. pp547-552, August-2018, Available at : http://www.jetir.org/papers/JETIR1808557.pdf