Abstract

This project focuses on semantic segmentation, the task of assigning a class label to each pixel in an image, using the Pyramid Scene Parsing Network (PSP Net). It involves training the PSP Net model and applying it to generate accurate predictions on new images. A key goal is to provide a user-friendly solution. To achieve this, a command-line interface (CLI) was developed, allowing users to easily train custom PSP Net models, create segmentation masks, and evaluate performance. The project leverages pre-trained PSP Net models on benchmark datasets like VOC12, ADE20K, and City scapes, making it a valuable tool for a variety of tasks such as object recognition and scene understanding. Its adaptability makes it a useful resource for both researchers and practitioners in the field of computer vision.