Contributor: Ashish Bangwal (ashish-2005)
Organization: OpenVINO Toolkit
Project: Industrial Meter Reading With OpenVINO (project page)
GitHub repository: ashish-2005/GSoC23-OpenVINO
The project is divided into two parts. In the first part of the project, we created a deep learning-based solution to read industrial meters. We achieved this by first detecting meters in the frame with the EfficientDet-d0 model, then doing semantic segmentation of meters with UNET (with a Resnet50 backbone) to obtain a segmentation map with which we calculated readings, and for inference we use the OpenVINO runtime. In part two of the project, we further optimize the solution pipeline to support multiple camera inputs with OpenVINO-toolkit functions like AsyncInferQueue and other performance hints, along with a web-based GUI with HTML,CSS, JS, and Flask web framework.
For further details visit project's Wiki page.
- Trained meter-detector and meter-segmentor along with their serialized OpenVINO Intermidiate-Representation.
- Notebooks demonstraings inference with Tensorflow and OpenVINO runtime.
- A web-app to monitor multiple meters and adjusting parameter to read meters with different range and intervals.
- Blog describing the whole work done in GSoC period.
- GSoC23 Contributor: Ashish Bangwal (ashish-2005)
- Mentor: Ethan Yang (OpenVINO-dev-contest)
- Mentor: Zhuo Wu (zhuo-yoyowz)
@ashish-2005 hii i m just beginner(my first contribution) and i wannna contribute to this organization pls guide me (i know python).