• Monisha K Naik
• Rekha B S

In this paper we look at Mask RCNN as a mechanism to automatically detect, classify and mask fish in underwater imagery. We describe the outcome of an analysis of a huge dataset of underwater pictures collected by the National Conservancy Organization using camera feeds gathered from boat cameras. The data is diversified; it lacks rotational symmetry across habitats, has particularly large shadows in comparison to the organisms under examination, and large occlusions and objects that are small and not centred in comparison to the whole field of view. The method provided here detects objects in an image efficiently while also producing a high-quality segmentation mask for each instance. Faster R-CNN is extended by Mask R-CNN, which adds a branch for predicting an object mask concurrently with the existing branch for bounding box recognition. Mask R-CNN uses the same two-stage technique, with the first stage being identical (which is RPN). Mask R-CNN outputs a binary mask for each RoI in the second stage, in addition to predicting the class and box offset. Despite the significant disparities in segmentation and classification outcomes when compared to land-based image datasets, the results are comparable to state-of-the-art efforts connected with land-based applications. The system provides effective detection and classification of fish using Mask RCNN with 90% classification accuracy. An automated model is developed by making use of train and test images of fishes in order to identify/ classify the fish species to particular class of its species using Mask RCNN algorithm.

Machine Learning, Object Detection, Convolutional Neural Network, Mask RCNN.

"Fish Detection and Classification using Mask R-CNN ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.8, Issue 6, page no.e19-e27, June-2021, Available :http://www.jetir.org/papers/JETIR2106556.pdf

"Fish Detection and Classification using Mask R-CNN ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.8, Issue 6, page no. ppe19-e27, June-2021, Available at : http://www.jetir.org/papers/JETIR2106556.pdf