We are seeking undergraduate and master’s students who are motivated to work on exciting projects within the LaSER lab. Below is the list of our available projects. If you are interested in volunteering to conduct relevant research or write a thesis on any of these projects, please feel free to reach out to Dr. Taeyeong Choi via email at email@example.com with your latest CV and a brief description of your previous experience and why your are interested in the particular topic.
If you are a (prospective) PhD student, you are also encouraged to contact Dr. Taeyeong Choi to discuss your research interests in detail.
“Realistic Image Synthesis by Prompting”
- There are generative models, designed for realistic image synthesis from text prompt inputs. For example, the image on the right shows the output of DALL-E2 given the prompt “Apple orchard with a robot”. This project is to leverage these methods to automatically generate a large dataset for training deep-learning vision models. Specifically, a Python-based interface will be implemented to interface with generative platforms and assess the synthesized outcomes. Through this project, you will gain insights into various text prompt-based image synthesis models, data acquisition processes, and methods for quality evaluation.
- Selective harvesting involves the process of identifying and gathering only healthy and ripe crops, and automation through robotics presents a promising solution. The aim of this project is to implement a Python-based interface between camera sensors and actuators of a robot arm will be implemented, enabling the robot to perceive and pick the fruits using robotic grippers. Participating in this project will provide you with hands-on experience in robot programming, perception, and motion planning.
“Simulator Development for Agricultural Active Robot Vision”
- Vision tasks within agricultural fields require robots capable of actively maneuvoring to find obstruction-free viewpoints. This project focuses on designing agricultural simulators for prototyping such robots, enabling them to navigate through realistic 3D orchard models. You will build these simulation frameworks in C++ and/or Python, leveraging existing platforms, like HELIOS, AgML, and DAVIS-Ag. Engaging in this project will provide you with an opportunity to delve into practical simulator design, a crucial step for robotics research. Successful deliverables are expected to have a significant impact on various agricultural robotics applications.
“Exploration of Image Data Augmentation Methods for Agricultural Applications”
- Data augmentation is critical for successful training of computer vision models using deep learning techniques to prevent overfitting issues but achieve reliable performance in real-world scenarios. This project aims to explore various image augmentation algorithms, including Channel Randomisation, analyzing their effects on agricultural tasks. The project will involve the development of novel methods tailored to the unique characteristics of the agricultural domain. Through this project, you will have the opportunity to gain hands-on experience in the development of deep learning pipelines for computer vision problems in Python, training and testing effective models for practical applications.