Classification-of-Gastrointestinal-Diseases-of-Stomach-from-WCE-using-the-improved-Saliency-based-me

Classification of Gastrointestinal Diseases of Stomach from WCE using the improved Saliency-based method and Color Features

Proposed Methodology

Segmentation

Active Contour-based segmentation

Algorithm 1: Initialize the Mask (φ^0) Step 1: Set the initial Mask to zero Step 2: MinPixelRange ← 50, MaxPixelRange ← 210 Step 3: rows ← MinPixelRange, cols ← MinPixelRange Step 4: For rows to MaxPixelRange if (rows == MinPixelRange) MaxPixelValue ← Selected Channel (rows, cols) End if
For cols to MaxPixelRange if (Selected Channel (rows, cols) ≥ MaxPixelValue) MaxPixelValue ← Selected Channel (rows, cols) R ← rows C ← cols End if End inner For End outer For Step 5: Create MaskValue1 ← ((R+C))/2, and MaskValue2 ← MaskValue1/2 Step 6: Get final Mask ((MaskValue2, MaskValue1), (MaskValue2, MaskValue1))

Improved Saliency-based segmentation

Figure 4: Improved saliency-based segmentation effects.

Figure 5: Proposed fusion effects using MAP approach.

Feature Extraction

Figure 6: Flow diagram of Features extraction and reduction.

Results

Segmentation Accuracy

Classification Accuracy

Confusion Matrix

Computation Time

Results Comparison & Analysis