01-07-2025
To help colourblind people see colours, Surat tech student develops technique
An MTech student of Sardar Vallabhbhai National Institute of Technology (SVNIT), Surat, has won an award for developing an image recolouring technique to help people with colour blindness. Saakshi Padmawar's research paper was selected for the best paper award at the OPJU International Technology Conference, held at O P Jindal University, Chhattisgarh, recently.
Padmawar's research focuses on improving the visual experience for individuals with Colour Vision Deficiency (CVD), popularly known as colour blindness, a condition that affects around eight per cent of the global male population and 0.4 per cent of women and girls.
People with CVD often find it difficult to distinguish between certain colour combinations, especially red and green, or blue and yellow, which can make it challenging to interpret traffic lights, educational charts, websites and more.
Padmawar, a student of Computer Science, developed an algorithm which can help people to see a wider range of colours. 'This can be used on the screens of gadgets. In the future, such techniques can be used in smart eyeglasses, allowing people to enjoy seeing different colours. I thought about helping people in the medical field. I chose the topic as it came under our subject project. I worked for five to six hours a day for five months on this topic,' she told The Indian Express.
According to Padmawar, 'By applying recolouration techniques, we aim to separate red and green hues effectively, enhancing distinguishing while maintaining visual naturalness. The ability to perceive colours arises from the complex interaction of light with the human visual system, particularly the retina cone cells. Any disruption in this process can lead to colour vision deficiencies that impact a person's daily activities and overall quality of life.'
The paper — titled 'Colourblind-Friendly Image Recolouring Using Generative Adversarial Networks' — was prepared under the guidance of associate professor Dr Chandra Prakash.
Sources said that out of 440 submissions from across the country to the OPJU International Technology Conference, her paper was chosen among the top 35 before it went on to receive the Best Paper Award.
The paper notes how colour blindness often goes unnoticed for years because people with the condition may adapt to their environment without realising that they perceive colours differently. This is especially true for mild to moderate forms of CVD, where individuals may only have trouble distinguishing specific colours, such as red and green or blue and yellow.
Human colour vision relies on three types of cone cells in the retina L-cones (sensitive to long wavelengths, associated with red), M-cones (sensitive to medium wavelengths, linked to green), and S-cones (sensitive to medium wavelengths, related to blue).
Dr Chandra Prakash said, 'In dichromatism, the absence of one type of cone limits colour perception…. Protanopia results from the absence of L-cones, while deuteranopia arises from the absence of M-cones. This leads to a significant overlap in perceived red and green hues, which diminishes contrast. For individuals with Protanopia or Deuteranopia, the red-green opponent channel is severely impaired.'
'Our study addresses these deficiencies, particularly for Protanopia and Deuteranopia. These conditions affect red-green differentiation, creating challenges in distinguishing colours in natural and artificial scenes,' Dr Prakash said.
Padmawar further said, '… Our methodology consists of three parts: simulating dichromatic vision, recolouring using linear transformation techniques, and generalising recolouration for unseen images using a GAN-based approach. Every phase is intended to address a different facet of improving dichromatic people's visual perception.'
