PrintingSegmentationToPrint

AI-based web application capable of segmenting an image into layers to be sent to print. The basic idea is to allow the user to give the possibility to upload an image, carry out the segmentation inference process, and the extraction of the layers in images. As well as being able to automatically detect, in case it is available, the option to carry out the inference on GPU, and if not, the default option, on CPU.

The ‘Printing Segmentation To Print’ project aims to develop an innovative web application that uses artificial intelligence (AI) to segment images into different layers, facilitating their subsequent printing. This tool will allow users to upload an image and, through an AI-based inference process, decompose it into multiple layers according to the elements present in the image. Each layer will be extracted as an independent image, ready to be sent for printing or processing according to the user’s needs.

The application will be designed to be highly efficient and adaptive, incorporating automatic hardware detection functionality. If the system detects the availability of a GPU, it will use this to speed up the segmentation process; otherwise, it will perform the inference using the default CPU. This ensures that the application works optimally in different environments, from high-performance machines to more modest devices.

The development of the project will be based on the use of technologies such as Python for programming, PyTorch for implementation, and web frameworks for the creation of an intuitive and user-friendly interface. In addition, ONNX will be used to optimise the models and ensure efficient performance on various platforms.

‘Printing Segmentation To Print is emerging as a practical solution for professionals and amateurs who need to accurately segment images for printing. With the ability to handle different hardware capabilities, this tool promises to be versatile and efficient, facilitating the workflow from segmentation to final print.


TOP