• Sujay S
• Siddagangu S
• H.R. Irfan Pasha
• Rekha B S

The goal of study is to create a "Fish detection and species categorization using Mask RCNN" model. Faster R-CNN has a variant-called Mask R-CNN. For object identification tasks, the faster R-CNN is commonly used. It returns the class name and bounding box coordinates of each object in the image. Mask R-CNN is simple to set up, and it only adds a little amount of overhead to Faster R-CNN. Mask R-CNN is divided into two phases. The First phase, it generates a premise about the possible locations for an object based on the image. Second phase, based on the primary state proposition, it predicts the object’s class, refines the bounding box, and creates a mask at a pixel level of the object dependent on the primary stage proposition. The system proposes the detection and species classification of the fishes AlbacoreTuna, BigEyeTuna, YellowFinTuna, MoonFish, DolphinFish, Shark. The system provides accuracy of 93%.

Object Detection, Region Proposal Network, Bounding Box, Mask R-CNN, Resnet101

"FISH DETECTION AND SPECIES CLASSIFICATION USING MASK RCNN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.8, Issue 7, page no.e507-e511, July-2021, Available :http://www.jetir.org/papers/JETIR2107568.pdf

"FISH DETECTION AND SPECIES CLASSIFICATION USING MASK RCNN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.8, Issue 7, page no. ppe507-e511, July-2021, Available at : http://www.jetir.org/papers/JETIR2107568.pdf