CHROMA is the new innovative solution that revolutionizes lighting in photography. Our project implements an inexpensive microcontroller system to identify the skin hue of a person and automatically adjusts environmental lighting in real time. By integrating advanced camera technologies, including auto-exposure and face detection, with AI-powered lighting solutions, CHROMA always ensures the best light for each subject. Whether for professional photography or live performances, CHROMA delivers best-in-class lighting adjustments for individual skin tone and makes high-quality photography increasingly available and more affordable.
Computer Engineer
Juan is a Computer Engineer major specializing in both hardware and software. As part of the CHROMA project, he is responsible for programming the microcontrollers and the Image Capture Unit (ICU), ensuring seamless integration and precise operation of the system. Fun Fact: he made this website ;)
Electrical Engineering
Nikul, an electrical engineer specializing in PCB design and power systems, is focused on developing automated lighting solutions that enhance accessibility and inclusivity for people of color.
Photonic Science and Engineering
Juanita specializes in optics and photonics, leading the optical design of the Image Capture Unit (ICU) for CHROMA, ensuring accurate representation of darker skin tones through innovative lighting solutions.
Photonic Science and Engineering
With a passion for live entertainment and over a decade of experience, Mads brings expertise in LED panel design to CHROMA, working to create optimal lighting for performers of all colors.
The optical design of our ICU comprises 4 important aspects: a large plano-convex lens to ensure enough light is entering the system, a plano-concave lens to magnify the image, an achromatic doublet to correct chromatic aberrations, and a 1/2.3" imaging sensor for accurate skin tone analysis. The housing comprises modular lens mounts and tubes to allow light to travel through the optical path only and focus on distances ranging from 6 to 12 feet. We have essentially created a zoom lens, but for a lower price than those on the market, more dynamic with the interchangeable tubes, and custom-made to fit CHROMA's needs.
The CHROMA system leverages computer vision and machine learning to intelligently adjust lighting based on a subject's skin tone. Using a Raspberry Pi with a PiCamera and the face_recognition and OpenCV libraries, our software detects faces in real-time and isolates the forehead region for precise skin tone analysis. These tones are classified using a custom-trained model based on the Monk Skin Tone Scale, enabling accurate mapping to one of five lighting presets. This classification is determined using color distance metrics over a calibrated dataset, emulating a machine learning decision process. Once the appropriate tone is identified, the system wirelessly communicates with two ESP32 microcontrollers via HTTP to activate synchronized lighting profiles using WLED. This closed-loop pipeline provides seamless lighting adjustments that are both inclusive and adaptive.
CHROMA's custom-designed main PCB is engineered specifically for precision control of five-channel RGBCCT LED arrays, enabling smooth and flicker-free dimming through hardware-level PWM. Unlike typical ARGB setups, our implementation leverages the ESP32's advanced PWM capabilities to drive high-power analog LEDs through WLED firmware-modified for non-addressable use. This approach offers a professional-grade lighting control solution at a fraction of the cost. A CP2102 USB-to-UART bridge is integrated to provide effortless firmware flashing and serial debugging, streamlining development and field updates. To ensure hardware protection, the board features a resettable polyfuse that guards against reverse polarity and overcurrent faults, safeguarding both LEDs and control electronics.
Power regulation in CHROMA is handled by a robust, dual-stage design optimized for reliability and noise suppression. At the input stage, a 24V supply-chosen for its efficiency in driving long LED strips-is stepped down using a Texas Instruments LM2576 buck converter to a regulated 5V rail. This efficient switch-mode regulator minimizes heat and ensures stable power under variable loads. The 5V output is then further refined by an AMS1117 linear regulator, delivering a clean and noise-free 3.3V line for sensitive digital components, including the ESP32-WROOM module. This linear stage adds an extra layer of filtering, critical for maintaining Wi-Fi stability and preventing brownouts during high-current transitions in LED modulation.
Standard lighting is often tuned for human vision, not cameras - resulting in images that appear natural to the eye but fail under camera sensors due to poor spectral distribution. Our custom-built LED panel addresses this challenge by delivering studio-quality illumination through a combination of 96 high-density chip-on-board (COB) LEDs, engineered specifically for photography applications. The panel integrates full-spectrum RGB diodes alongside dual high-CRI white channels, enabling highly accurate color rendering and adaptable lighting conditions for any subject. This hybrid approach offers an unprecedented level of spectral control, critical for maintaining consistent tone and exposure across varied skin types. The panel supports precise correlated color temperature (CCT) tuning, ranging from a warm 3200K tungsten to a daylight-balanced 6500K, and is capable of dynamic brightness modulation via wireless control. Designed with thermal efficiency, modularity, and color harmony in mind, our lightbox ensures that subjects are lit with fidelity—whether in a controlled environment or a live performance setting.
| Senior Design I | Senior Design II |
|---|---|