When a painter, violinist, and clarinetist join forces, the possibilities for creative collaboration brim with potential. Of course, it doesn’t hurt that the painter, Felicia Iordachi, also works with software as a Math and CIS major. Then there’s the fact that violinist Elise Choe and clarinetist Amrah Rasheed work in hardware as EE majors. Together, their potential for creative collaboration translated to tangible results: The Poetry Camera, which won the best overall final project prize for ESE 1110 Atoms, Bits, Circuits, and Systems.
In conceptualizing the project, they wanted to capture the beauty of everyday moments either missed or overlooked by enterprising Penn undergrads. As a handheld device, the Poetry Camera is powered by a Rasberry-P-5 that captures any image into a generative poem. In pressing the shutter button, the camera takes a photo, analyzes the scene using a camera module, sending the extracted features to an AI language model, explains Choe, Iordachi, and Rasheed. Within seconds, the device prints an original poem on thermal receipt paper using an onboard thermal printer. In total, the Poetry camera comes with eleven different modes of poems, including Haiku’s and Limerick’s. As for setting the poem’s tone? The Poetry Camera utilizes environmental awareness to analyze visual cues such as shadows, the number of people in the room and empty seats, the team explained.
Akin to the “musicality” of their final project, the group’s composition lent itself to seamless collaboration. “We worked well together,” Rasheed said, explaining that as a CIS major, Iordachi had expertise handling the camera’s software components. Rasheed and Choe in turn had the background to handle its hardware design. “It was a very natural progression,” Rasheed emphasized.
Speaking of natural progression, Iordachi championed Choe’s ability to keep the team on track. “She was very organized and acted as the project manager.”
Despite the group’s balance of skills, they nonetheless encountered challenges with heat management, power issues, the camera’s physical design, and GPS limitations.
According to the team, the thermal printer generated significant heat, galvanizing them to install a fan to prevent melting. When it came to supplying power, they tested several batteries until they found one powerful enough to provide enough current for the thermal printer and both microcontrollers. Regarding its size, the original compact CAD design needed to be upscaled three times to accommodate the wires and airflow. As for its GPS limitations, the GPS module failed inside the case. This led the team to utilize AI to determine location as a result.
Overcoming the above challenges led to late night sessions in the Detkin lab, including a 36-hour sprint before the deadline. Nonetheless, the group thrived from the lab’s energetic atmosphere, describing the lab at night as an inspirational source of energy. There, the group was surrounded by fellow freshmen as well as seniors working with robots for senior design.