Estimation of Fish Pose with Human Structure using Multi Level Neural Network

Abstract

Pose estimation refers to the method of estimating human positions from a photo. Pose estimation can be done both in 3-D and in 2-D which helps to analyse the motion. During the early days of human pose estimation, classical techniques called picture structures had been furnished. To come to be aware of human beings successfully in snapshots, key factors at the body are mainly positioned to determine their pose. Human hobby, popularity, human- monitoring, laptop interplay, gaming, signal- languages, and video surveillance all required to estimate the positions. It's been proposed in this paper that several procedures are used to resolve this trouble. XGBoost Gradient neural network algorithm is experimented for the human pose dataset and shows better result for fish pose estimation.