patent2

Technical Field of the Invention

Medicine and Ophthalmology: This innovation is designed and implemented to improve the accuracy and performance of physicians and ophthalmologists, with clear applications in this field.
Biomedical Engineering: This idea utilizes biomedical engineering sciences to enhance and support diagnostic and therapeutic methods, serving as a bridge between engineering and medicine for design optimization.
Electrical and Computer Engineering: The device incorporates electronics and artificial intelligence from computer engineering in various system components.
Mechanical Engineering: Mechanical engineering principles are applied in designing the body, casing, covers, and system molds.

Technical Problem and Invention Objectives

Vision is one of the most important special senses, not only because it plays a crucial role in connecting us with the external environment, but also because visual impairment can significantly reduce learning capabilities. While diseases like diabetes and cardiovascular conditions have the highest prevalence rates, eye diseases remain among the most significant health issues affecting many individuals.
Many eye diseases, if left untreated (especially during childhood), can lead to blindness and cause irreversible problems, creating numerous limitations in patients’ personal and social lives. Therefore, early identification and diagnosis of diseases has always been crucial. This invention aims to achieve early detection of such conditions while emphasizing preventive strategies to identify diseases and their underlying risk factors.

Understanding Heterochromia

Heterochromia is a term describing differences in iris color within an individual. Central heterochromia occurs when different colors appear in the same eye, while complete heterochromia is when each eye has a different color. This condition results from variations in melanin concentration and distribution—the pigment responsible for skin, hair, and eye color.
Eye color is determined by melanin storage in the iris, which controls pupil dilation to regulate light entry. Blue eyes have minimal melanin, while brown eyes are melanin-rich.
The iris, the most visible part of the human eye, originates from the choroid layer and consists of blood vessels and connective tissue. Melanocytes produce melanin pigment, and other pigmented cells determine eye color. The iris contains smooth muscles that control light passage by adjusting pupil size—the opening in the iris center. This process regulates light reaching the retina.

Anatomical Structure

The iris comprises two distinct layers with different embryonic origins:

• Anterior margin: Composed of stroma—a loose layer containing connective tissue with melanocytes and non-pigmented cells
• Posterior surface: Covered by densely packed pigmented epithelial cells forming two tightly connected layers

Eye color differences result from pigmentation levels in the deep stromal layer, primarily the anterior margin and stromal density, controlling light absorption and reflection from the iris.

Types and Causes

Heterochromia can indicate other underlying conditions. Early detection during infancy and childhood can be crucial. Associated conditions include:

• Bloch-Sulzberger syndrome
• Bourneville disease
• Hirschsprung disease
• Horner’s syndrome
• Parry-Romberg syndrome
• Sturge-Weber syndrome
• von Recklinghausen disease
• Waardenburg syndrome

Acquired heterochromia develops later in life due to disease, injury, or medication, while congenital heterochromia is present from birth or has genetic origins.
Central heterochromia features two different colors in one iris—typically an outer ring of one color and an inner ring of another, often called “cat eyes.” The outer color is considered the true eye color and occurs more frequently in irises with low melanin levels.
Sectoral heterochromia (partial heterochromia) appears as an irregular spot in the iris rather than a ring around the pupil.

Innovation Objectives

This innovation develops a precise system using engineering sciences to:

1. Enable early detection of eye diseases, focusing on heterochromia in infancy
2. Identify underlying diseases and pathogenic factors associated with this condition
3. Utilize AI algorithms to improve diagnosis and reduce human error
4. Provide completely non-invasive diagnosis and identification
5. Monitor all eye parameters and diseases related to size, color, and uniformity

By achieving early detection with high accuracy during infancy, this system helps treat underlying conditions and prevents the progression and complications arising from undiagnosed heterochromia.

Prior Art and Background of Related Advancements

Significant efforts and progress have been made in retinal image processing to develop automated systems for diagnosing various diseases. These systems not only enable processing large volumes of retinal images with minimal time and cost, but also eliminate fatigue and other human limitations that diagnosticians may experience.
Recent work has focused on automatic analysis and diagnosis of conditions such as diabetic retinopathy, age-related macular degeneration, and premature retinopathy. Research continues to develop new algorithms capable of processing images with varying quality and illumination while minimizing errors. Additionally, image processing techniques and quantitative measurement of retinal blood vessel topography are being used to study the relationship between retinal microvessels and cardiac diseases.
The ability to transmit images remotely has increased the use of image processing techniques in clinical decision-making for retinal images from rural medical centers—now known as Tele-Ophthalmology. Taking images at regular intervals and recording them enables studying pattern changes and disease progression stages on the retina.

Advanced Imaging Technologies

The new 3D-OCT retinal imaging technique provides high-quality, high-precision data and images from different retinal layers. This has reduced algorithm errors in automatic diagnosis and extraction of retinal disease patterns, leading to novel algorithms for automatic image analysis.
Research groups have utilized multi-scale curvelet transform in image processing to extract diabetic retinopathy patterns (exudates, hemorrhages, microaneurysms) along with optic disc location and retinal blood vessels from digital color retinal images. Work on OCT images and AMD disease detection is ongoing.

Major Retinal Diseases

Two major retinal diseases are glaucoma and diabetic retinopathy:

• Glaucoma is the second leading cause of blindness worldwide, characterized by optic nerve cupping and visual field loss
• Diabetic retinopathy can lead to blindness but can be prevented through early detection and annual examinations

Retinal Imaging Methods

Common retinal imaging methods include:

• Ophthalmoscope (introduced by Hermann von Helmholtz in 1851)
• Fluorescent angiography (introduced by Novotny and Alvis in 1961)
• OCT (introduced by Huang in 1991)
• Newer methods such as ultrasound and laser blood flow measurement

OCT imaging can capture structural information, blood flow, elastic parameters, polarization changes, and molecular content. Using wavelength measurement principles through light interference, it creates 2D or 3D high-resolution images of anatomical cross-sections.

OCT Image Processing Approaches

The primary focus in OCT image processing is segmentation in 2D or 3D. Despite OCT technology evolving since 1991, data segmentation has only emerged in the past decade as one of the most challenging and necessary steps in analysis. No universal segmentation method suitable for all applications has been developed.
Processing approaches can be categorized into five types:

1. A-scan applicable methods: First introduced by Hee, prevalent until 2005, but had limitations including long processing times and low layer detection accuracy
2. B-scan applicable methods: Reduced 2D noise through preprocessing but required complex, time-consuming noise reduction techniques like anisotropic diffusion
3. Active contour methods: First proposed by Cabrera Fernández and optimized by Yazdanpanah, showing better noise resistance than intensity-based methods
4. AI-based analysis methods: Using support vector machines or fuzzy c-means clustering techniques, with processing times around 45 seconds and 2-pixel error rates
5. 3D graph-based methods: The most suitable approach, achieving processing times of 45 seconds for 3D volumes (480×512×128 voxels) with high accuracy (2.8 micrometer error)

Current Limitations and Innovation Need

Most existing systems are designed for retinal examination, while others focus on the cornea or lens. These systems generally:

• Lack adequate precision
• Contain human error potential
• Are not specialized for iris examination

This has driven the development of our specialized system that:

• Enables early detection capabilities
• Is specifically designed for iris studies
• Reduces human error through advanced algorithms
• Provides high-precision measurements for iris-related conditions

The innovation addresses the critical need for a dedicated iris examination system with early detection capabilities, filling a gap in current ophthalmic diagnostic technology.