研究目的
To create awareness about breast cancer and develop a noninvasive technique using statistical features from mammogram images to detect and classify normal, benign, and malignant cases for early diagnosis.
研究成果
The research demonstrates that statistical features extracted from mammogram images using ImageJ can effectively differentiate between normal, benign, and malignant cases, aiding in early cancer detection. This noninvasive technique emphasizes morphological differences and shows promise for reducing mortality through automated detection, with suggestions for future work to enhance accuracy and apply to larger datasets.
研究不足
The study relies on a specific database (DDSM) which may limit generalizability; the image processing techniques might be sensitive to noise and variations in image quality; and the statistical approach may not capture all morphological aspects of cancer, potentially requiring more advanced methods for higher accuracy.
1：Experimental Design and Method Selection:
The study uses a cascade algorithm based on statistical parameters (mean, median, standard deviation, perimeter, skewness) to process mammogram images for classification. Image processing techniques include preprocessing (conversion to grayscale or RGB formats using ImageJ), image sharpening (high boost filtering, high-frequency emphasis), contrast enhancement (normalization, histogram stretching, histogram equalization), background subtraction, and edge detection.
2：Sample Selection and Data Sources:
Mammogram images for normal, benign, and malignant cases are obtained from the Digital Database for Screening Mammography (DDSM) database.
3：List of Experimental Equipment and Materials:
ImageJ software is used for image processing and feature extraction.
4：Experimental Procedures and Operational Workflow:
Steps include preprocessing images to specific formats, sharpening to reduce noise and highlight edges, enhancing contrast through normalization and histogram equalization, subtracting background to isolate tissues, extracting statistical features, and detecting edges using filters.
5：Data Analysis Methods:
Statistical parameters are calculated and compared across image types to classify them based on prescribed value ranges.
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
