ProjectsCase studyOffline

Object Detection Using TensorFlow

Real-time object detection with deep convolutional neural networks — the practical system behind a published research paper.

Role

Co-author and developer

Timeline

2019

Stack

Python / TensorFlow / OpenCV

Year

2026

Object Detection Using TensorFlow

Fig. 01 — Object Detection Using TensorFlow

(01)

The problem

Object detection research is easy to publish and hard to run. The paper needed a system that actually detected objects in a live stream, not results from a fixed test set.

(02)

The approach

Implemented real-time detection over deep convolutional neural networks with TensorFlow, using OpenCV for the capture and annotation pipeline, and measured it against the paper's evaluation.

Real-time
Live video inference
2019
Paper published
(03)

The outcome

Published as a research paper during the final year of my Computer Science degree.

Built with

01PythonIntermediate
02TensorFlowIntermediate
03OpenCVIntermediate