SIT740 - Research and Development in IT: Object Detection Project

This project investigates the application of object detection techniques, specifically using TensorFlow, to enable robots to recognize real-world objects. The core issue addressed is how to equip robots with the ability to identify objects such as faces, bicycles, and buildings. The solution involves utilizing the TensorFlow object detection API to build a system capable of recognizing objects. The project highlights the advantages of this approach, such as enabling the deployment of robots for commercial purposes. The limitations include the fact that the robots can only recognize objects predefined in their database. The project also discusses the potential for creating more intelligent robots that can differentiate between valuable and ordinary objects. The project includes screenshots of the simulation and references relevant research papers.

Running head: RDIT - DISTINCTION TASK 8.3
SIT740 Research and Development in Information Technology
Distinction Task 8.3: Investigation of tool for a given problem
Student Name:
University Name:

1RDIT - DISTINCTION TASK 8.3
Recognition of real time world objects with Object detection
The issue that is rising in context to the advancement in robotics is how to teach the robots to
recognize world objects such as faces, bicycles, and buildings using object detection. This is
considered as the major concern in this particular study. It has been determined that the identified
problem can be solved with the help of Object Detection technique comprising of suitable
algorithms. The tool that has been chosen for this study is TensorFlow which is an open-source
software library used for dataflow programming. The main purpose of this study is to resolve the
existing issues with teaching the robots on how to detect and recognize real world objects. This
can be done with the help of implementing the Object Detection API designed for execution in
TensorFlow. The major advantage of the solution is that it will help to deploy the robots for
commercial purposes. The drawback associated with the solution is that the robot will be able to
only recognize the objects that are predefined in the database. The attractive feature of the
solution is that it will help to create a new level of intelligence in the robots so that they can
easily differentiate between the valuable and ordinary objects.

2RDIT - DISTINCTION TASK 8.3
Screenshots of the simulation running related to the issue
Figure 1: Design of the program related to the problem
(Source: Created by Author)
Figure 2: Detection of objects in the image
(Source: Created by Author)

3RDIT - DISTINCTION TASK 8.3
Bibliography
Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G.S., Davis, A.,
Dean, J., Devin, M. and Ghemawat, S., 2016. Tensorflow: Large-scale machine learning on
heterogeneous distributed systems. arXiv preprint arXiv:1603.04467.
Shrivastava, A., Sukthankar, R., Malik, J. and Gupta, A., 2016. Beyond skip connections: Top-
down modulation for object detection. arXiv preprint arXiv:1612.06851.