研究目的
Design and operate a computer-aided decision system for optimal choosing the best refractive surgery based on patient needs, starting from corneal topography and aberration images, using RGB and HSI color spaces and decision tree.
研究成果
A software for this Computer-Aided Diagnosis system was developed by the authors. The graphical interface is a user-friendly one. Starting from classical Corneal topography and aberation images, using a decision tree which is a technique of classification after obtaining the RGB and HSI values, the Computer-Aided detect system chooses refractive surgery, percentage of vision correction, ablation and residual stroma with high precision. Images are stored in databases together with patient personal details. The software includes image processing modules, databases and decision tree.
研究不足
The accuracy of the system was 71.06, which may increase by adding new color image processing.
1：Experimental Design and Method Selection:
The study involved 124 patients (248 eyes) intending to undergo refractive surgery by excimer laser. Patients with general conditions such as diabetes or collagen disease were excluded. The study used RGB and HSI color spaces and decision tree for image processing and classification.
2：Sample Selection and Data Sources:
Patients had a routine refractive surgery screening and examination including the determination of uncorrected visual acuity, refraction, topography using the Orbscan II corneal topographer (Bausch and Lomb) and aberrations using a wavefront sensor (Zywave, Bausch and Lomb).
3：List of Experimental Equipment and Materials:
Orbscan II corneal topographer (Bausch and Lomb), wavefront sensor (Zywave, Bausch and Lomb).
4：Experimental Procedures and Operational Workflow:
The system analyzed the image of the patient’s corneal topography and aberration and recognized them, offering a response that contains the best four possible diagnoses, ordered by a better result, with the percentage of vision correction, ablation and residual stroma for each of the ordered refractive surgery.
5：Data Analysis Methods:
The classification was conducted using the decision tree in the attempt of determining the minimum identification rule by means of if-then rules.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
