Dalsilens: A Real-Time Mobile Assistive Tool for Dichromacy Using LMS Daltonization and HSV-Based Color Identification

Abstract

Color Vision Deficiency (CVD), commonly known as color blindness, is a genetic condition that affects a person’s capacity ability to differentiate between certain light wavelengths. Although eyesight itself is usually normal, individuals with CVD find it difficult to distinguish colors within the red, green, and blue spectrums. The most common type is Dichromacy, which includes Protanopia (red-blind), Deuteranopia (green-blind), and Tritanopia (blue-blind). Existing assistive technologies, such as color-filtering glasses, are designed to separate overlapping red and green wavelengths, help with color distinction but are often expensive. Additionally, many existing smartphone applications require users to upload images for processing, which prevents real-time operation and limits immediate visual feedback. This paper proposes the implementation of the LMS Daltonization algorithm on the Android platform using real-time image processing. The method converts the camera’s input from the RGB color space to the LMS color space. This allows for the mathematical simulation and adjustment of the missing color spectrums. The application is expected to help users distinguish objects with specific color spectrum through a smartphone, with low latency that supports real-time usage.

Why this research

Color plays a big role in everyday tasks, but for users with dichromacy, it can be unreliable and tiring to interpret. Dalsilens starts from that problem and tries to make colors easier to understand in real-life situations.

Approach

The current approach combines two parts that work together:

1. LMS daltonization to shift red, green, and blue so their differences are easier to see.
2. HSV-based identification to show a basic color name when pointing at a specific pixel.

What we’re testing

• Whether shifting red, green, and blue through LMS daltonization actually makes color differences easier to notice in real-time.
• How useful basic color labels are when users point at specific parts of a scene.
• Whether the system can stay responsive enough to be used continuously without feeling laggy.