project-logo
ZGen Tech
November 2017 - June 2018
Junior Software Engineer
Home

An advanced object detection system designed to not only classify objects within images but also to autonomously search the web for related information, enriching the understanding of the visual content.

project-image
project-image
project-image

An Introduction

What began as an internal investigation into advanced tools for enabling computers to identify and classify objects within scanned images—and autonomously search the web for related content—was deceptively simple in concept. The challenge, however, was to make it work on modest devices like a Raspberry Pi while providing near-instant results. This project not only achieved its objectives but also became the catalyst for my software development career, setting the stage for a path rich in innovation and exploration.

Skill Breakdown

[ Frontend ]

15%

[ Backend ]

89%

[ Database ]

64%

[ Other ]

24%

Major Technologies and Concepts Used

[ Frontend ]

< Frontend development skills were not a necessity for the completion of this project. />

[ Backend ]

skill-logo

Python

skill-logo

PyTorch

skill-logo

Tensor Flow

[ Database ]

skill-logo

SQL

[ Other ]

skill-logo

Git

skill-logo

Github

Enhanced object recognition capabilities paired with intuitive understanding.

My initial foray into software development was nothing short of fascinating. What began as a straightforward task to assist in developing an object recognition tool quickly escalated into an intense journey of research, architectural design, and ultimately, the creation of a sophisticated solution. This tool was not merely capable of detecting various objects but also classifying them and conducting web searches based on those classifications, showcasing the depth and breadth of its capabilities.

 

An innovative object detection system tailored for computers, showcasing the remarkable skill to autonomously roam and search the internet.

- Yahffa Jagrup

 

The research phase of the project was a deep dive that culminated in the discovery of two pivotal tools perfectly suited for our needs: TensorFlow and PyTorch. Each presented its own set of challenges in terms of usability and integration, yet both emerged as invaluable assets for their robust object classification capabilities.

 

By integrating custom logic to extract statistics from the classified objects, we laid a solid foundation that seamlessly enabled the use of this information for web searches, enhancing the system's functionality and intelligence.

 

Harnessing the strengths of PyTorch and TensorFlow, two prominent forces in machine learning, I propelled the research efforts toward groundbreaking innovation.

- Yahffa Jagrup

 

The implementation phase of the project was truly captivating, marking my inaugural foray into Python's versatile applications. It was an opportunity to architect and develop a scalable solution, meticulously engineered to enhance performance through the adoption of asynchronous techniques and parallel computing strategies. This approach significantly optimized both image processing and web searching components, showcasing the project's innovative and efficient design.

 

Integrating parallel computing techniques into the program ignited my passion for software development, turning a routine task into an exhilarating challenge.

- Yahffa Jagrup

Topics That Captured My Interest

Explore the diverse topics and insights that have captured my curiosity.

Delving into the strengths and weaknesses of both TensorFlow and PyTorch and ultimately choosing to develop the application with TensorFlow proved to be a wise decision that paid dividends in the later stages of development.

This project underscored the realization that development largely focuses on the creation and transformation of data into diverse formats.

A practical hands-on experience with Python profoundly deepened my knowledge of real-world applications and strategies for solving problems in a professional context.

Final Thoughts

Undertaking this project was a transformative journey that broadened my technical horizons and deepened my understanding of the intricate dynamics between data creation, transformation, and application. Through exploring object recognition with TensorFlow and PyTorch, I leveraged TensorFlow's strengths to bring the ask into reality—an exemplar of efficient information curation rather than just a tool. Transitioning from concept to a robust assistant, this project highlighted the transformative potential of software development, blending Python proficiency with challenges in networking, API development, and parallel computing. This experience marked a pivotal chapter in my career, providing profound insights, surmounting challenges, and laying a solid foundation for future innovations at the nexus of technology, creativity, and determination.

Similar Showcases